Revisiting Observed Changes in Cloud Properties over Europe

Självständigt arbete Nr 34

Revisiting Observed Changes in Cloud Properties over Europe

Revisiting Observed Changes in Cloud Properties over Europe

Patrik Boström

Patrik Boström

Uppsala universitet, Institutionen för geovetenskaper Kandidatexamen i fysik, 180 hp

Examensarbete C i meteorologi, 15 hp Tryckt hos Institutionen för geovetenskaper Geotryckeriet, Uppsala universitet, Uppsala, 2012.

The Earth’s atmosphere is a vulnerable system which is easily changed by micro- and macrophysical variations. Big decreases in pollution levels of sulfur dioxide over Central Europe from 1980s to 2000s led to decreased mass concentration of atmospheric solid and liquid particles. This gives the opportunity to investigate how these particles influence the atmosphere. Newly released satellite climatology data was used to analyze statistics of cloud properties during four years in the high polluted atmosphere (1985-88) and four years in the less polluted atmosphere (2004-07). These two periods were investigated in collaboration with Atmospheric Remote Sensing Unit of the research department of the Swedish Meteorological and Hydrological Institute (SMHI). Cloud top temperature of liquid clouds over polluted regions during the earlier period was colder by more than 2 K and more than 5 K for only optical thin liquid clouds. The changes in mass concentrations of atmospheric particles derived by the sulfur dioxide emissions are shown to be a highly possible factor to the observed cloud changes.

Självständigt arbete Nr 34

Revisiting Observed Changes in Cloud Properties over Europe

Patrik Boström

Supervisors:

Abhay Devasthale, Atmospheric Remote Sensing,

Research Department, Swedish Meteorological and Hydrological Institute

Monica Mårtensson, Air, Water and Landscape Sciences, Uppsala University

2

3

A

R

E

The Earth’s atmosphere is a vulnerable system which is easily changed by micro- and

macrophysical variations. Big decreases in pollution levels of sulfur dioxide over Central Europe from 1980s to 2000s led to decreased mass concentration of atmospheric solid and liquid particles. This gives the opportunity to investigate how these particles influence the atmosphere.

Newly released satellite climatology data was used to analyze statistics of cloud properties during four years in the high polluted atmosphere (1985-88) and four years in the less polluted

atmosphere (2004-07). These two periods were investigated in collaboration with Atmospheric Remote Sensing Unit of the research department of the Swedish Meteorological and

Hydrological Institute (SMHI). Cloud top temperature of liquid clouds over polluted regions during the earlier period was colder by more than 2 K and more than 5 K for only optical thin liquid clouds. The changes in mass concentrations of atmospheric particles derived by the sulfur dioxide emissions are shown to be a highly possible factor to the observed cloud changes.

4

R

O

E

Patrik Boström

Jordens atmosfär är ett känsligt system som lätt förändras av mikro- samt makrofysikaliska variationer. Stora minskningar i föroreningsnivåer av svaveldioxid över centrala Europa från 1980 till 2000-talet ledde till minskade masskoncentrationer av fasta och flytande atmosfäriska

partiklar. Detta ger en möjlighet att undersöka hur dessa partiklar påverkar atmosfären. Nyligen utvecklad klimatologisk satellitdata användes för att analysera statistik av molnegenskaper under fyra år i en högt förorenad atmosfär (1985-88) och fyra år i en mindre förorenad atmosfär (2004- 07). De två perioderna undersöktes i samarbete med Enheten för atmosfärisk fjärranalys av forskningsavdelningen till Sveriges meteorologiska och hydrologiska institut (SMHI).

Molntopptemperaturen för moln i vätskefas över förorenande områden under den tidigare perioden var mer än 2 K kallare och mer än 5 K kallare för endast optiskt tunna moln i vätskefas.

Förändringarna i masskoncentrationer för atmosfäriska partiklar och droppar med svaveldioxidusläpp som ursprung visas vara högst möjliga att ligga bakom de observerade molnförändringarna.

5

C

Abstract ... 3

Revisiting observed changes in cloud properties over Europe ... 3

Referat ... 4

Omprövning av observerade förändringar i molnegenskaper över Europa ... 4

1 - Introduction ... 7

1.1 - Earth’s climate system ... 7

1.1.1 - Observed dimming and brightening ... 8

1.2 - Clouds and aerosols in the climate system ... 8

1.2.1 - Pointing out aerosols ... 9

1.2.2 - The role of clouds ... 9

1.2.3 - The role of aerosols ... 10

1.2.4 - Aerosol-cloud interactions ... 10

1.2.5 - An overview of SO2 emissions over Europe ... 13

1.2.6 - Main objective of the present study ... 14

2 - Methods, data processing and analysis ... 16

2.1 - Calculations... 17

3 - Results ... 18

3.1 - Changes in cloud-top temperatures over Europe ... 18

3.1.1 - Clouds with optical thickness greater than 3.6 ... 18

3.1.2 - Sensitivity to optical thickness ... 19

3.1.3 - Non-precipitating clouds ... 23

3.1.4 - Number of observations ... 24

3.1.5 - Summary of the results ... 25

4 - Discussion ... 26

4.1 - Analyzing the changes in CTT ... 26

4.1.1 - Possible factors changing CTT ... 26

4.1.2 - Using discounting principle for attribution of changes ... 27

4.1.3 - Sensible C1 clouds ... 30

4.1.4 - Connecting CTT to SO2 emissions ... 30

4.2 - Odd CTT changes ... 31

6

4.2.1 - CTT increases over North and Baltic sea ... 31

4.2.2 - Number of observations ... 31

4.3 - Future work ... 31

5 - Conclusions ... 32

6 - Acknowledgements ... 33

7 - References ... 34

Appendix A - List of abbreviations ... 38

Appendix B ... 39

7

1 - I

A list of all abbreviations used in the report is shown and described in Appendix A.

1.1 - E

’

The climate is defined as statistics of conditions from long term measurements (typically 30 years) of the Earth’s inner systems at a given region. The inner systems are: the atmosphere (meteorological conditions), hydrosphere (the hydrologic cycle), cryosphere (part of a planet covered by ice), geosphere (the dense inside of Earth) and biosphere (all life systems on the planet) [Bridgman and Oliver, 2006]. Each of these components of the climate system is intrinsically connected to each other and in addition impacts one another. A change in one condition that changes another which in turn changes the first condition is called feedback [Le Treut, Somerville et al., 2007]. The atmosphere will be of interest in this work and thus, the climate conditions are referring to statistics of meteorological variables.

The sun is the climate system’s main source of energy. The solar radiation interacts with earth and its components. The relationship between earth’s incoming (short wave) and outgoing radiation (long wave) in watts per square meter (W m^-2) is called the Earth’s radiation budget (ERB).

Incoming radiation primarily interacts with the atmosphere where it is absorbed, scattered or reflected. The surface solar radiation (SSR) is the sum of all solar radiation (in W m^-2) that passes through the atmosphere and interacts with the Earth’s surface [Wild, 2009]. By this, SSR is a driving force to the ERB. The Earth’s atmosphere balances incoming energy partly by reflecting and partly by emitting long-wave thermal radiation [Le Treut, Somerville et al., 2007]. A typical balance of average radiation distributed over Earth is shown in fig. 1.

Fig. 1 Estimates of Earth’s main radiation exchanges. All units are in W m^-2. These estimates show that approximately 235 W m^-2 of the incoming solar radiation stays in the Earth’s systems as distribution of long wave radiation [Source: Kiehl and Trenbert, 1997; (c)American Meteorological Society. Reprinted

with permission., www.ametsoc.org].

8

About 30% of all solar radiation in a temporal and spatial average over Earth is reflected back to space. This is called the planetary albedo (def. Albedo = Reflected radiation / Incident radiation, in %) and is caused by reflection of atmospheric particles, cloud, ground etc. [Ackerman and Knox, 2007]. To determine the SSR we also have to investigate the atmosphere’s optical thickness (also known as optical depth). The optical thickness is a measure of a medium’s transparency to radiation [Ackerman and Knox, 2007]. For example, a cloud-filled atmosphere has a greater optical thickness than a cloud-free atmosphere since it lets less solar radiation pass through it.

A change of SSR impacts the ERB which in turn implies a climate change. To change the Earth’s amount of SSR, either an internal or external variation has to occur. External variations are primarily changes in Earth’s orbit and the solar output since other extraterrestrial effects (e.g.

activities of other celestial bodies) are in general negligible. A change in Earth’s inner systems is internal and can be a change of the atmosphere’s optical thickness since it is essential for the amount of incident SSR [Le Treut, Somerville et al., 2007]. An increase or decrease in a factor that forces the ERB to change is called radiative forcing (RF) and is measured in the same unit as ERB and SSR (i.e. W m^-2) [Wild, 2009]. Positive RF leads to an increase in total energy and hence a positive effect on the ERB, called brightening. The opposite effect is dimming (negative RF).

1.1.1 - OBSERVED DI MMING AND BRI GHTENING

Observations have shown significant changes in SSR over big parts of the Earth during the 20^th century and studies show that Europe follows these trends too. Measurements of SSR from stations located worldwide have been recorded in Global Energy Balance Archive (GEBA) at the Swiss Federal Institute of Technology, Zurich and maintained by the World Radiation Data Center located at The Main Geophysical Observatory in St. Petersburg [Gilgen et al., 1998].

Data of global monthly average SSR from GEBA during the period 1960 to 1990 show a distinct dimming with a total decrease in SSR of approximately – 7 W m^-2 (– 2.3 W m^-2 decade^-1) during this period [Wild, 2009]. The opposite variation in SSR was found for the following years after the dimming. Wild et al. (2005) estimated an increase in SSR of 6.6 Wm^-2 decade^-1 based on an averaged value from 1992 to 2002. Limiting the area to Europe, Norris and Wild (2007) showed the same trends for this continent solely.

To study the dimming and brightening period, an analysis of the solar radiation’s path through the atmosphere is needed to be done.

1.2 - C

Aerosols are defined as a suspension of solid and/or liquid particles in a gaseous medium which implies a system of two or three phases. Atmospheric aerosols consist of air and airborne solid and/or liquid particles [Ackerman and Knox, 2007].

Atmospheric aerosols can be launched into the air by nature itself (natural aerosols) or they may have a human made source (anthropogenic aerosols). The longevity of atmospheric aerosol particles varies from few hours to several weeks. One exception is aerosol particles positioned far above the highest clouds, they can stay airborne for one to two years [Leaitch et al., 1999].

9

Aerosol particles may enter the atmosphere by being launched directly into the air, these are known as primary aerosols. Common primary aerosols such as sea salt and dust are thrown up in the air by sea sprays (bubble bursting) and the wind. Secondary aerosols enter the atmosphere by:

(1) in situ (at the spot) transformation from gas to particle called nucleation [Kulmala et al., 2004], (2) condensation on a particle [SMHI, 2012].

Sulfates (e.g. SO₄) are secondary aerosol particles derived from both anthropogenic and natural emissions of the gas sulfur dioxide, SO₂. The emitted SO₂ mixes with water droplets by

condensation and form H₂SO₄, then the evaporation of the polluted droplets leave the sulfates SO₄ alone. This is the most common forming of SO₄ but the other usual occurring forming is direct nucleation of SO₂ [SMHI, 2012]. Volcanic activity emits SO₂ to the atmosphere and is the major natural producer of SO₄ aerosol. However, fossil fuel burning emits almost 5 times as much SO₂ making this anthropogenic source the main source of SO₄ aerosol [Penner et al., 2001].

1.2.1 - POINTING O UT AERO SO LS

In order to determine the origin of the 20^th century’s dimming and brightening periods we investigate which parameters changes the SSR and compare these changes with the observations mentioned above. Remember, the dimming period (1960-1990) had a decrease in SSR of

approximately –2.3 W m^-2 decade^-1 and the brightening period (1992-2002) showed increased SSR of approximately 6.6 W m^-2 decade^-1.

Studying the Earth’s orbital cycles, one detects that they varies in 10,000 to 100,000 years which implies too long periods to have notable effects over a decade. The sun’s radiation during the 20^th century varied in a sunspot cycle but only with an increase of 0.17 W m^-2 decade^-1. Therefore, variations of solar output can be neglected as a source to dimming or brightening [Willson and Mordvinov, 2003]. This leads to that the factors in debt for the variations in SSR are internal to Earth and hence the atmosphere’s optical thickness will be studied. Changes in optical thickness can be developed from changes in properties and amount of clouds, water vapor and finally aerosols.

Water vapor covers from 0 to 4 % (depending on temperature etc.) of the total composition of the atmosphere. Nonetheless, studies show that the global total amount water vapor has increased with 4% from 1960 to 2000 due to an average increase in temperature but the SSR decrease from this effect is merely –0.1 W m^-2 decade^-1 [Wild et al., 2007]. This excludes water vapor as well to be a crucial factor filling the observed changes in SSR. Summing the RF from all atmospheric gases gives –0.31 W m^-2 decade^-1 [Wild, 2009].

Derived above, the possible parameters to change the SSR as observed are changes in clouds and aerosols. In order to find the “missing” RF we have to analyze clouds, aerosols and their effects.

1.2.2 - THE RO LE O F CLO UDS

To the mass and volume, clouds mainly consist of water droplets or frozen crystals. Due to their in general high albedo and covering 60% of the Earth’s surface, clouds play an important role in the ERB. Common summer clouds such as shallow cumulus and stratocumulus reflect a big

10

portion of incoming solar radiation and absorb or scatter the remaining part. Clouds absorb and emit radiation back to the atmosphere. Despite keeping parts of the long wave radiation in the atmosphere, clouds have a cooling effect because of their high albedo (see

fig. 1) [Le Treut et al., 2007].

An increase in total cloud amount implies a dimming effect (negative RF) based on the cooling from clouds. An increase in cloud reflectivity of approximately 5% would hide the entire increase in greenhouse gases (GHGs) during the industrial era [Voiland, 2010]. Changes in other macro physical properties such as altitude and structure also contribute to RFs. Looking at the smaller scale, a minor change in micro physical properties of clouds (e.g. droplet size, phase) will affect the ERB distinctively as well because they affect radiation characteristics and feedbacks the macro physical properties [Collins et al., 1994].

1.2.3 - THE RO LE O F AERO SO LS

As ever-present, the aerosol particles impact the SSR. The aerosol RFs were categorized by IPCC’s (Intergovernmental Panel on Climate Change) Third assessment report: Climate change (known as TAR, 2001) into the direct, indirect and semi-direct effect. Furthermore the indirect effect contains subcategories [Forster et al., 2007].

The forming of a cloud starts with condensation of water vapor when the air is oversaturated. A cloud droplet is created by condensation on a particle surface. A portion of the atmospheric aerosol particles attract water and hence act as a cloud condensation nuclei (CCN) upon which water condense at a specific saturation and form cloud droplets. At lower temperatures, frozen crystals are formed by deposition of water vapor or freezing of liquid droplets on aerosol particles which then are called ice nuclei (IN) [Wild, 2009]. Both primary and secondary aerosol particles are able to act as a CCN or IN if they attract water.

The direct effect entails the scattering and absorbing/emitting of solar and thermal radiation from aerosol particles in the clear-sky atmosphere. The total RF from direct scattering and absorbing/emitting effect is negative, thus the result is a dimming at the Earth surface [Forster et al., 2007]. Important terms while studying aerosol direct effects are amount, chemical

composition and optical thickness of the particles. The optical thickness depends on the size, amount and optical properties of aerosols and implies how much of the incoming radiation will be scattered or absorbed. Both anthropogenic and natural aerosol emissions govern the direct effect [Wild, 2009].

The semi-direct and indirect effects are results from aerosol particle interactions with clouds; they will be defined in the following chapter.

1.2.4 - AERO SOL-CLOUD I NTERACTIONS

The aerosol semi-direct and indirect effects play an important role in the ERB. Since aerosol particles act as CCN or IN upon which cloud droplets or crystals form, different amounts and properties of the aerosols changes the amount and properties of clouds [Wild, 2009]. This leads to the indirect effect which in turn is divided into subcategories. The characteristics of the

11

aerosols and atmospheric conditions determine the occurrence of the indirect effects. See fig.2 below for an illustration of the indirect effects and the semi-direct effect.

By increasing the amount of aerosol particles acting as CCN, the number of cloud droplets or frozen crystals can increase. While having the same amount of water vapor available and increasing the CCNs, the droplet or crystal size will decrease which develops denser clouds. A denser cloud has a smoother surface which implies an increase in reflectivity (a higher albedo) giving this 1^st indirect effect the name cloud albedo effect [Lohmann and Feichter, 2005].

In case of liquid phase clouds, a cloud can either disappear by precipitation of its contents or being burned off by the sun. The precipitation of a cloud occurs when the cloud droplet radius exceeds the critical value and the weight of the droplets or crystals overcomes the updraft carrying the cloud. As mentioned above, increasing the aerosol particle number concentration will decrease the size of the cloud droplets or crystals and giving them a lesser weight and thus, the precipitation is delayed. In this way the cloud will have a longer lifetime. This is the 2^nd indirect effect and is called cloud life time effect [Haywood and Boucher, 2000].

Absorbing aerosols may cause an increase in absorption of radiation in the cloud layers and rise the temperature at their altitude. Through the higher temperature inside the cloud layers, the clouds will evaporate and being burned off by the sun faster. This effect is called the semi-direct effect and is caused primarily by black carbon [Lohmann and Feichter, 2005]. If the aerosols position are not inside but above the clouds they will absorb radiation before it reaches the cloud and thus lets it last longer. Both of the different variants of the semi-direct effect can be seen in fig. 2 [Denman et al., 2007].

An increase in the number of aerosol particles acting as INs can increase the frequency of super cooled clouds’ glaciation. This indirect aerosol effect is called the glaciation effect and has the ability to turn a cloud with water droplets into a cloud consisting of frozen crystals. Therefore, the glaciation effect can transform a non-precipitating cloud into one that precipitates due to the growth of frozen crystals in an oversaturated surrounding [Denman et al., 2007].

Smaller cloud droplets are more resistant to freezing than larger ones. Khain et al. (2001) pointed out and showed that only if the cloud droplets were small enough and many in number could they reach down to –37.5°C without freezing. Increased aerosols giving more and smaller cloud droplets would therefore imply that clouds reach colder temperatures before they glaciates. Also, more latent heat is released when the more numerous cloud droplets freeze and thus, a boosting of the internal updrafts occurs. There’s still only vague knowledge about this aerosol indirect effect which is called the thermodynamic effect. The thermodynamic effect is expected to give, higher clouds, lower cloud top temperatures (CTTs) and have a negative RF [Denman, et al., 2007].

The studies by Devasthale et al. (2005) and Koren et al. (2005) provide observational evidence of cloud thermodynamic effect based on satellite data.

12

Fig. 2 Aerosol indirect and semi- direct effects. TOA – Top Of the Atmosphere. Note that cloud albedo and cloud lifetime effects are shown in the same part of the figure, first in a clean cloud with no effects and then in a polluted cloud with both effects.

The same applies to the glaciation and thermodynamic effects in the lower part of the figure. This shows that the effects are not independent but provide feedbacks and influences each other [Source: Denman et al., 2007; Public figure from Climate Change 2007: The Physical Science Basis;

Working Group I Contribution to the Fourth Assessment Report of the

Intergovernmental Panel on Climate Change, Figure 7.20.

Cambridge University Press].

Studies of forest fires and ship contrails verify that an increase of aerosol particle number

decreases the size of cloud droplets. Experiments have shown the same relationships by emitting SO₄ as primary aerosol particles to act as a CCN [Forster et al., 2007]. This verifies SO₂ as an aerosol contributor. The experiments also estimated the average short wave radiation passing through these clouds to be decreased with –1.5 Wm^-2 which verifies the cloud albedo effect (1^st indirect effect) [Andronache et al., 1999].

The cloud albedo effect as well as the cloud life time effect reduces the solar radiation at the top of the atmosphere and hence reduces the SSR which implies a negative RF (dimming) from both effects [Lohmann and Feichter, 2005]. The estimates of the RFs by the indirect effects varies and contain uncertainties, however Ramaswamy et al. (2001) estimated a global mean RF from the

13

cloud albedo effect to be –1.2±0.7 W m^-2 in global mean values. Similar global values for the cloud life time effect were estimated to be –0.85±0.55 W m^-2 [Lohmann and Feichter, 2005].

The semi-direct effect gives both positive and negative RF depending on the position of the aerosols as seen in fig. 2. For absorbing aerosols inside the cloud layers, less radiation will be reflected and thus, a positive RF will occur. Absorbing aerosols above the clouds “steal” the radiation from the clouds and therefore leads this variant of the semi-direct effect to a negative RF [Denman et al., 2007].

It’s well-known that the total RF from all aerosol effects (direct and indirect effects) is dimming but still, the amount of RF is only uncertain estimations. The RF from all aerosol effects on the average SSR is estimated by climate models to be – 2.3±1 W m^-2 [Denman et al., 2007].

Simulations show that SSR is more vulnerable to changes in the atmosphere’s optical thickness than changes in GHGs [Denman et al., 2007]. From above, it’s also understood that the aerosol direct and indirect effects impact the atmosphere’s optical thickness which now shows the aerosols importance to the ERB.

1.2.5 - AN OVERVI EW OF SO2 EMI SSIO NS OVER EURO PE

Anthropogenic SO₂ originates from fossil fuel burning and from combustion of other sulfide substances. Continued industrial development in Europe during the 20^th century tossed aerosols into the atmosphere by increased fossil fuel burning and other polluting emissions. This trend sustained until 1980 when anthropogenic SO₂ emissions started to decrease in Western Europe.

The annual decrease of SO₂ emissions over Europe reached its maximum after the fall of the Eastern Bloc in 1989 (refers to former Soviet Union and countries of the Warsaw Pact).

Improvements in politics and infrastructures implied major decreases of pollutions over Central and Eastern Europe as well [European Commission, 2012; Vestreng et al., 2007].

Fig. 3 below shows the distinct reduction in SO₂ emissions over Central Europe. The Black Triangle is the heavily polluted area were Germany, Poland and Czech Republic meet. This figure also shows that the percentage decreases of emissions in the black triangle countries are greatest in Czech Republic.

14

Fig. 3 Percentage ratio change in SO₂ emissions in 2004-07 with respect to 1985-88 [Source: EDGAR, 2012; The data is public and obtained from European Database for Global Atmospheric Research

(EDGAR)].

The observed SSR changes over Europe as discussed above are matching well to the changes in SO₂ emissions. The decrease in SO₂ emissions are in line with the increasing SSR after 1990 and gives another reason why effects occurred by SO₂ should be investigated more. This leads to a suspicion of SO₂ as a potential factor in debt for the dimming. The fact that SO₂ is transformed from a gas into an aerosol particle and the aerosol effects gives aerosols also a possible role in the SSR variations. Additionally, Berglen et al. (2005) showed a reduction of SO₄ concentration in line with the reduction of SO₂ emissions.

1.2.6 - MAI N O BJECTIVE O F THE PRESENT STUDY

In this work, the statistical changes in cloud properties over Central Europe will be investigated in relation to the changes in SO₂ emissions using the newly released CM-SAF (Satellite Application Facility for Climate Monitoring) long-term cloud climatology data from European Organization for the Exploitation of Meteorological Satellites (EUMETSAT). This cloud climatology is based on Advanced Very High Resolution Radiometer (AVHRR) sensors onboard National Oceanic and Atmospheric Administration (NOAA) satellites.

Cloud microphysical and physical properties will be investigated (e.g. cloud top temperature, cloud optical thickness etc.). The CM-SAF data is used to analyze aerosol particles impact on cloud properties. The different SO₂ amounts in the atmosphere between the dimming and the brightening period can enlighten factors caused by changes in anthropogenic sources of SO₂. Statistics of cloud properties for the years 1985, -86, -87 and -88 will be analyzed with respect to

Ratio of change in SO2 emissions (%)

15

the years 2004, -05, -06, -07. Only days of the summer months (i.e. June, July and August; 92 days) will be used because of the seasonal variations in atmospheric conditions. The European area of interest is enclosed by the geographic coordinates 35°N to 60°N and 10°W to 35°E.

The main goal of the present study is to analyze changes in cloud properties, mainly cloud top temperature (CTT; temperature at the top of the clouds) and cloud optical thickness (COT; the optical thickness of clouds), between two periods and find the reasons behind the changes. The studies of SO₂ transformations into aerosol particles, the changed SO₂ emissions and the different SO₄ concentrations lets the number of aerosol particles to be assumed more in the 80s period than in the 00s. In this way, an attempt to verify the aerosol indirect effect called the thermodynamic effect will be made.

The present study derives its motivation from the study by Devasthale et al. (2005), wherein they examined changes in CTTs over the Central Europe also using AVHRR data. They found that cloud tops were higher during the late 80s compared to the late 90s and attributed this to the change in SO₂ emissions during these two periods.

The present study extends and improves their study in the following ways.

a) A different and extensively validated cloud property retrieval algorithm is used (Information available on http://www.cmsaf.eu and http://www.nwcsaf.org).

b) The state-of-the-art calibration of AVHRR channels is performed.

c) Even cleaner period, in terms of anthropogenic emissions of SO₂ over Europe, is selected in the early 2000s.

d) The study region covers a larger area.

e) Discounting principle is used to attribute changes in CTTs [Hull, J. G., and West. S. G., 1982].

16

2 - M

,

In the present study, the data from two AVHRR sensors onboard two NOAA satellites is used.

NOAA-9 is used for the period 1985-88 and NOAA-16 is used for 2004-07. AVHRR is a five channel instrument, with two channels in the shortwave wavelength spectrum, two in the thermal (split-window, centered at 11.0 and 12.0 µm) and the fifth one falls partly in the visible and partly in the thermal (centered at 3.7 µm wavelength). The radiances from the solar channels are

intercalibrated to remove any inconsistencies [Heidinger et al., 2010].

Based on AVHRR data, CM-SAF provides retrievals of cloud properties for the period beginning from 1982 until present [Schulz et al., 2009]. The cloud retrieval algorithm derives its heritage from the Nowcasting SAF’s Polar Processing System (PPS) software [Dybbroe et al., 2005a; Dybbroe et al., 2005b], which is technically under continuous development and validation. However, it is matured enough to allow climate studies involving investigations of the large-scale statistics as done in the present study. The retrievals of cloud properties (both physical and microphysical) have undergone rigorous validations, especially, during the past few years [Karlsson et al., 2008;

Roebeling et al. 2003; 2006; and Roebeling, 2008]. Further detailed documentation, algorithm theoretical basis documents and validation reports are available on the following two websites:

http://www.cmsaf.eu http://www.nwcsaf.org

CM-SAF’s level 2b data was used for the present study. This data set is available at 0.1 x 0.1 degree spatial resolution and twice daily (day and night). However, only daytime estimates were used since the additionally used microphysical retrievals (cloud phase and optical thickness) are available only during daytime. 92 daily observations / year were analyzed.

The methodology used in the analysis is shown below.

a) Focus is only on liquid water clouds using phase flag provided in the CM-SAF data.

b) We analyze clouds with COT greater than 3.6 and further partition data based on optical thickness of clouds into three subcategories to investigate sensitivity. The first category is clouds with optical thickness in the range of 3.6-9.4, the second category with optical thickness 9.4-23.0 and the third with optical thickness greater than 23.0. These optical thickness thresholds are selected to conform to the International Satellite Cloud Climatology Project (ISCCP) cloud type definition. We are essentially covering stratus, stratocumulus and shallow convection regimes.

Note that very deep convection will automatically be excluded since it is most likely to have ice tops. The COT-categories are not well-defined for a specific cloud type since different clouds can technically have the same optical thickness. However, the thinnest clouds may be still developing and the thicker clouds are not developing in the same grade.

c) Precipitation flag available in the data are used to exclude clouds with precipitation.

17

2.1 - CALCULATIONS

Average CTTs are calculated for the periods exclusively and then the difference between them is found. Statistics of all liquid clouds (both precipitating and non-precipitating) with COT greater than 3.6 and every COT subcategory will be separated from each other. Also, a category

including all liquid clouds with greater COT than 3.6 but with only non-precipitating clouds will be included.

The name of all COT categories and associated abbreviations are listed below:

All liquid clouds with COT ≥ 3.6 AC

Liquid clouds with 3.6 ≤ COT ≤ 9.4 C1

Liquid clouds with 9.4 < COT ≤ 23 C2

Liquid clouds with COT > 23 C3

Liquid clouds with no precipitation and COT ≥ 3.6 CNP

In addition to the mean value, the coefficient of variation (CV) is calculated. The deviation in CTT for every combination of COT cannot only be calculated with the standard deviation. The different prerequisites (i.e. different amount of SO₂ emissions) of the two periods lead to that the deviations are of different magnitude. To assess the deviation we therefore need a unitless value.

CV is defined as the standard deviation divided by the mean value and thus gives the rate of deviation from the mean value (i.e. unitless). In this way we can compare the CV of CTT for the different clouds and periods independent to the different prerequisites [Alm and Britton, 2008].

18

3 - R

3.1 - C

-

E

The first objective in this study was to investigate if the CTTs have changed over Europe during two time periods; one with more polluted conditions (1985-88) and the other with less pollution (2004-07).

Note: The values of the southern latitudes (e.g. Black and Mediterranean Seas) as well as of Baltic and North seas in fig. 4-10 will be explained in the

discussion (Ch. 4.2).

3.1.1 - CLO UDS WI TH OPTICAL T HI CKNESS GREATER THA N 3.6

The first result (fig. 4) is the difference between the 00s’ and 80s’ average CTTs for AC clouds.

The results show a distinct rise of the CTT for AC clouds from the 80s to 00s period. Over Central Europe and also over United Kingdom we can see an increase of the CTT ranging from 0 to 6 Kelvin. Large increases in CTT for AC clouds can be seen almost all over Europe.

Fig. 4 Difference in CTT (Kelvin) for AC between 00s and 80s. A positive value means an increase of the CTT to the 00s period.

CTT (Kelvin)

19

3.1.2 - SENSITIVITY TO OPTI CA L THICKNESS

By limiting the COT between 3.6 and 9.4 (C1) we observe even clearer increases. Fig. 5 shows increases up to about 10 K which means that C1 clouds have CTTs increasing even more than the average of all clouds (fig. 4). CTT for C1 clouds not only have a high maximum increase but also a high average increase over big parts of Europe.

Fig. 5 Difference in CTT (Kelvin) for C1 clouds between 00s and 80s.

CTT (Kelvin)

20

The clouds with 9.4 < COT ≤ 23 (C2) have also increased CTTs but compared to both AC and C1 clouds they have a less maximum CTT and a much less spatial average CTT. As seen in fig. 6, the CTT over the Central Europe are from just negative to 2 K positive in average. C2 clouds still have local increases in CTT but overall the increase is not significant.

Fig. 6 Difference in CTT (Kelvin) for C2 clouds between 00s and 80s.

CTT (Kelvin)

21

The results of the C3 clouds show an increase of CTT as well (fig. 7). CTT of these clouds are greater than for C2 clouds but still much less than for the thin C1 clouds. The C3 clouds over center of figure have an increased CTT for up to about 6 K but 2K in average over Central Europe.

Fig. 7 Difference in CTT (Kelvin) for C3 clouds between 00s and 80s.

CTT (Kelvin)

22

The calculated CV values show that the CTT varies more during the 80s than the 00s. The most interesting CV values can be seen in the C1 clouds (fig. 8 and 9). These values differ much between the periods and show the greatest CV, especially over polluted land regions. Thus, the CTT of C1 clouds varies most of all clouds. The CVs of C3 is a little greater than C2 clouds but overall both types stay approximately unchanged between the two time periods and are therefore not shown here. The CVs for all COT categories for both periods are shown in Appendix B (fig.

B1).

Fig. 8 CV of CTT for C1 clouds for 1985-88.

Fig. 9 CV of CTT for C1 clouds for 2004-07.

CVCV

23

3.1.3 - NO N-PRECIPITATING CLO UDS

The CNP clouds for all COTs (fig. 10) follow nearly the same trends as for AC. CNP clouds’

CTT are in average approximately the same as AC clouds’ CTT.

Fig. 10 Difference in CTT (Kelvin) for CNP clouds between 00s and 80s.

CTT (Kelvin)

24

3.1.4 - NUMBER O F O BSERVATION S

Around 30 out of 92 days during the 80s period had clouds following the AC restrictions given above and categorized to corresponding COT group. The 00s period had 4±1 days less. Center of fig. 11 and 12 shows a large number of correct observations and near zero at southern latitudes.

Fig. 11 Total number of cloudy observations for 1985-88.

Fig. 12 Total number of cloudy observations for 2004-07.

Number of cloudy observations Number of cloudy observations

25

3.1.5 - SUMMARY O F THE RESULTS

A short list of all results from the calculations is shown below.

Increase in CTT: The CTTs of liquid clouds over Europe are higher in 00s period compared to 80s period.

The deviating thin clouds: The CTTs of C1 clouds have increased more than the other liquid clouds.

Coefficient of variation: The C1 clouds show the biggest coefficient of variation in CTT between 00s and 80s.

Non-precipitating clouds: The CTTs of CNP clouds follow the same trends as AC.

Number of cloudy observations: The number of cloudy observations has slightly

decreased over Central Europe and remains unchanged at southern latitudes.

26

4 - D

4.1 - A

CTT

The study of CM-SAF long term satellite level 2b data have shown lower CTTs of liquid clouds over land in 1985-88 than in 2004-07. We’ve also seen the unique difference in CTT for optical thin clouds with 3.6 ≤ COT ≤ 9.4 compared to the other cloud categories. The first objective of this study, i.e. to investigate changes in CTTs, has been achieved. The question is now: what are the reasons behind the changes?

First, we try to figure out what different CTTs means to the properties of clouds. Since the temperature in average decreases with the height, it means that a higher cloud top has a lower temperature. The observed differences in CTTs may not only be because of changed cloud top heights but the differences are too big to exclude changed heights as an influencing factor. This means that the cloud tops are higher in the more polluted period. Then why are the cloud tops higher in the 80s compared to the 00s?

Above, we’ve seen the changes in SSR and pointed out aerosols to be the most probable factor.

While studying the changes in CTTs, further work is needed to point out aerosols as a potential factor to these changes as well. Implementations of atmospheric computer models to study the behavior of an atmosphere with and without added anthropogenic aerosols would be a way to find what has caused the changes. However, the limit of time in this study will not allow any studies of this grade to be implemented. When not using models we need to look as many factors as possible that could cause changed CTTs using observational data. By analyzing every factor we try to give reasons why a given factor should not cause changes in CTTs using so-called

discounting principle. In this way, all factors will be eliminated until hopefully only one is left.

4.1.1 - PO SSI BLE FACTORS CHANG ING CTT

Main factors that could contribute to changes in CTTs are 1) changes in atmospheric circulation, 2) changes in surface temperatures and amount water vapor, 3) satellite calibration errors and 4) changes in aerosol properties. Explanations to every factor are listed below.

1) Changes in atmospheric circulation: The changes in winds and circulation have direct impacts on the cloud distribution and macro physical properties. A change of the average wind direction can for example lead to an increased transportation of clouds from land towards sea where the conditions are completely different.

2) Changes in surface temperatures: Since the analysis covers only summer seasons, the surface temperature plays an important role. A warmer surface implies more convection and hence increases the cloud top heights.

27

3) Satellite calibration and retrieval errors:

Errors in data from the satellites may originate from errors in calibrations. The difference in the two satellites deviation from the orbit can lead to errors as well.

4) Changes in aerosol particles: The atmospheric aerosol particles affect the

atmosphere as shown in the introducing chapter of this work, in fig. 2. They may through the indirect effects also contribute to changes in CTT.

4.1.2 - USING DI SCOUNTING PRI NCIPLE FO R ATTRI BUTI O N OF CHANGES

Every of the abovementioned factors will here be shown why it should or should not be responsible for the observed changes in CTT.

1) Changes in atmospheric circulation: The average wind speed and direction for 850 hPa level over the same two four-year periods (1985-88 and 2004-07) are shown in fig. 13 and 14 below [ECMWF, 2012]. The figures are plots obtained from European Centre for Medium-Range Weather Forecasts’ (ECMWF) ERA-Interim project (additional information available on http://www.ecmwf.int/). The wind direction is nearly unchanged which means that the clouds transport in approximately the same direction. The wind speed has decreased slightly (less than 1-2 m/s), but we believe that this should not be sufficient to cause such large changes in CTTs. Thus, changes in atmospheric circulation can be excluded as a primary cause.

28 Wind speed (m/s)

Wind speed (m/s)

Fig. 13 Average wind speed (m/s) and direction from 1985-88 of summer months [Source: ECMWF, 2012; Follows License conditions for ECMWF archive products].

Fig. 14 Average wind speed (m/s) and direction from 2004-07 of summer months [Source: ECMWF, 2012; Follows License conditions for ECMWF archive products].

29

2) Change in surface temperatures: The ground surface temperatures over Europe are warmer in 00s compared to 80s and also the total amount water vapor has increased. This should lead to higher cloud tops in 00s via potentially increased convective activity as well as available moisture. But instead, the opposite was found, the cloud tops were higher in 80s. The C1 clouds can be clouds under development and should also get a higher cloud top from increased convection. Therefore, we can rule out this factor well.

3) Satellite calibration and retrieval errors:

The differences in brightness temperatures between AVHRRs onboard NOAA-9 and NOAA-16 are less than 1 K. We have additionally focused on daytime data where the retrieval accuracy is higher. Level 2b data is prepared in such a way that only the best estimates and retrievals close to nadir are used. Time periods and satellites are chosen in such a way that the orbital drift of these satellites is similar (fig. 15).

Therefore, the CTT changes due to errors in satellite data are minimal.

Fig. 15 Day-time equatorial crossing time for NOAA polar satellites

[Source: STAR, 2012; With permission from and developed by Dr. Felix Kogan and Dr. Wei Guo at Center for Satellite Application and Research (STAR), NOOA, United States of America].

4) Changes in aerosols: That leaves us with the other important factor which has major influence on clouds, i.e. aerosols. There are many studies establishing that increased aerosol solid

30

and liquid particles can lift the cloud tops via

invigoration resulting from the additional latent heat release. The aerosol particle and droplet (number and mass) concentration was most probably higher in 80s compared to 00s as the SO₂ was higher. Thus, SO₂ is most likely to be one of the factors changing the clouds as observed.

4.1.3 - SENSIBLE C1 CLO UDS

The most significant increases that the results showed were for the C1 clouds. The differences between the two periods are bigger for these clouds than the others. In addition, C1 clouds show a great ratio of deviation from the mean value (CV) in the aerosol-filled 80s and a much smaller in the 00s. By comparing the difference of CV between 80s and 00s period for all cloud types we see that the thin liquid clouds can be more sensible to at least long-term changes in SO₂ pollution levels. Since the C1 clouds may either be developing clouds or liquid water limited regimes, they will through the indirect effects be more sensitive to aerosol particle changes which fit well to our observations and the changed SO₂ emissions.

C2 clouds are most likely to be fully developed stable systems, wherein relatively speaking, aerosol changes probably have second order influence. Therefore, C2 clouds do not show the same variations as C1 clouds. We focus only on liquid water clouds why the optical thickest C3 clouds can be shallow convective clouds and not deep convective clouds (containing ice). The convective movements in the C3 clouds should lead to that these clouds are susceptible to aerosol particle changes. C3 clouds should therefore vary little more between the periods than C2. This is consistent with our observations and C3 clouds are therefore more sensible to the aerosol particle changes than C2 clouds.

4.1.4 - CONNECTING CTT TO SO2 EMISSIONS

The changed CTTs observed in this study are both temporal and spatial in line with the changed anthropogenic emissions of SO₂; the CTTs are lower when more anthropogenic SO₂ are present.

The aerosol particles derived by the SO₂ emissions are most likely affecting the cloud properties.

However, the thermodynamic effect cannot fully be proved in this study. It is partly because we only know that the mass of SO₄ aerosols is greater in the earlier period and not the aerosol particle number. Also, focus has only been on investigating the CTT and COT using the CM- SAF data.

31

4.2 - O

CTT

4.2.1 - CTT I NCREASES O VER NO RTH AND BALTI C SEA

Figures 4-7 and 11 show an increase in CTT over the North and Baltic Sea. As the SO₂ emissions does not transform into aerosol particles direct, the wind will carry the pollutions from the industries to adjacent areas before they turn into aerosols.

The average wind-direction (fig. 13 and 14) is westerly over United Kingdom and west- southwesterly over Central Europe. The wind will therefore in average carry pollutions from United Kingdom to the North Sea and from Central Europe to the Baltic Sea. Thus, the changes in CTT over these two seas originate from changes in industrial activities.

4.2.2 - NUMBER O F O BSERVATION S

The decreased number of observations during the 00s period over the center of our studied area (Germany etc.) seen in fig. 11 and 12 should not cause uncertainties. The number of cloudy observations in 00s is still one third of all observations. However, the southern latitudes have in both periods too few observations to give a fair value. Therefore we see the patchiness and big variations over the southern latitudes (e.g. Mediterranean and Black Seas), these values can be neglected.

4.3 - F

This study covers the changed amount of anthropogenic SO₂ emissions and the changed mass of SO₄ in the atmosphere. The theory of an acting thermodynamic effect could be enhanced if we could prove that also the number of aerosol particles is higher in the earlier period.

A different area with similar changes in anthropogenic SO₂ emissions can be studied and implement a comparison to the present study. An example of other interesting areas is China where we’ve seen an industrial development in different periods.

More factors and more independent data should also be investigated. For example, only SO₂ emissions have been considered and not the changes in black carbon emissions or other pollutions that could lead to aerosol particles. Land based observations of clouds could be included to also investigate how more than the top of the clouds have changed.

In this study we have to content ourselves with seeing that the changes in CTT follow nearly the same periods as the SSR and that the aerosols probably are the factor causing the changes. The amount RF from a colder and higher cloud top should be investigated further since the observed effect is most likely to be a factor changing the climate.

32

5 - C

Clouds and aerosols play an important role in the Earth’s radiation budget. The massive

infrastructural and political changes in Europe during the 80s and 90s lead to significant changes in pollution loads and thus probably changes in aerosol particle number concentrations before and after the late 80s. This provides a unique opportunity to study corresponding changes in cloud properties and investigate the possible role of aerosols in influencing these changes via indirect effects. The newly processed, extensively validated and accurately calibrated cloud data set from the CM-SAF allowed us to investigate such changes in cloud properties.

Since the main focus of the present study was on studying the possible aerosol impact on cloud thermodynamics, we focused on investigating changes in cloud top temperatures of liquid clouds over Europe during the 2004-07 (cleaner period) compared to 1985-88 (highly polluted period).

In line with the previous study by Devasthale et al. (2005), we also find that cloud top temperatures were lower during the late 90s compared to early 00s. Additionally, we find that the CTTs of thin liquid clouds (3.6 ≤ COT ≤ 9.4) were changed the most. These clouds also show a greater deviation from the mean value compared to other clouds during the periods. Several parameters that can potentially change CTTs were investigated such as atmospheric circulation, surface temperatures, errors in satellite data etc. However, none of these factors can convincingly explain the observed changes in CTTs and only the substantial changes in SO₂ emissions and aerosol regimes during the selected time periods are large enough to have an impact on cloud

thermodynamics.

In future, we will investigate even more factors that could influence CTTs and also use other independent data sets to examine similar statistics.

33

6 - A

I would like to acknowledge the entire CM-SAF Cloud product team for providing cloud retrievals for my study. I would also like to thank SMHI and Uppsala University for their support. Special thanks to Abhay Devasthale and Monica Mårtensson for their help completing this work.

34

7 - R

Ackerman, S. A. and Knox, J.A. (2007). “Meteorology: Understanding the Atmosphere”. Thomson Brooks/Cole. pp.48-50

Alm, S. E. and Britton, T. (2008). “Stokastik: Sannolikhetsteori och statistikteori med tillämpningar”.

Liber AB. pp. 63-67

Andronache, C., Donner, L. J., Seman, C. J., Ramaswamy, V., and Hemler, R. S. (1999).

“Atmospheric sulfur and deep convective clouds in tropical Pacific: A model study”. J. Geophys. Res., 104. pp.

4005– 4024.

Berglen, T. F., Myhre, G., Isaksen, I. S. A., Vestreng, V., and Smith, S. (2007) “Sulphate trends in Europe: Are we able to model the recent observed decrease?”. Tellus. doi:j.1600-0889.2007.00289.x 1111/j.1600-0889.2007.00289.x, 2007.

Bernes, C. (2007). “En ännu varmare värld: Växthuseffekten och klimatets förändringar”. Fälth & Hässler, Värnamo, Sweden. pp.20-21.

Bridgman, H. A. and Oliver J. E. (2006). “The Global Climate System: Patterns, Processes and Teleconnections”, p.1. Cambridge university press

Collins, W. D., Conant, W. C., and Ramanathan, V. (1994) “Earth radiation budget, clouds, and climate sensitivity”, The Chemistry of the Atmosphere: Its Impact on Global Change, edited by: J. G.

Calvert, Blackwell Scientific Publishers, Oxford, UK. pp. 207–215

Denman, K.L., Brasseur, G., Chidthaisong, A. Ciais, P., Cox, P.M., Dickinson, R.E.,

Hauglustaine, D., Heinze, C., Holland, E., Jacob, D., Lohmann, U., Ramachandran, S., da Silva Dias, P.L., Wofsy, S.C. and Zhang, X. (2007). Couplings Between Changes in the Climate System and Biogeochemistry. In: ”Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change”. [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M.Tignor and H.L. Miller (eds.)].

Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. pp.555-567 Devasthale, A., O. Krüger, and H. Grassl. (2005). “Change in cloud top temperatures over Europe”.

IEEE Geosci. Remote Sens. Lett., 2(3). pp.333–336

Dybbroe, A., Thoss, A., and Karlsson, K.-G. (2005a). “NWCSAF AVHRR cloud detection and analysis using dynamic thresholds and radiative transfer modeling – Part I: Algorithm description”. J. Appl.

Meteor., 44- pp. 39–54,

Dybbroe, A., Thoss, A., and Karlsson. (2005b). “K.-G.: NWCSAF AVHRR cloud detection and analysis using dynamic thresholds and radiative transfer modeling - Part II: Tuning and validation”. J. Appl.

Meteor., 44. pp.55–71,

European Centre for Medium-Range Weather Forecasts (ECMWF). (2012).

http://www.ecmwf.int/research/era/do/get/era-interim

35

European Commission. (2012).

http://ec.europa.eu/ireland/about_the_eu/fall_of_eastern_bloc/index_en.htm European Database for Global Atmospheric Research (EDGAR). (2012).

http://edgar.jrc.ec.europa.eu/background.php

Forster, P., Ramaswamy, V., Artaxo, P., Berntsen, T., Betts, R., Fahey, D.W., Haywood, J., Lean, J., Lowe, D.C., Myhre, G., Nganga, J., Prinn, R., Raga, G., Schulz, M. and Van Dorland, R.

(2007). Changes in Atmospheric Constituents and in Radiative Forcing. in: “Climate Change 2007:

The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the

Intergovernmental Panel on Climate Change”. [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M.Tignor and H.L. Miller (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. pp.153-180

Gilgen, H., Wild, M. and Ohmura, A. (1998). “Means and trends of shortwave irradiance at the surface estimated from global energy balance archive data”, J. Climate, 11, 2042–2061.

Haywood, J., and Boucher O. (2000). “Estimates of the direct and indirect radiative forcing due to tropospheric aerosols: a review”. Reviews of Geophysics, 37. Pp. 513-543

Heidinger, Andrew K.; Straka, William C. III; Molling, Christine C.; Sullivan, Jerry T. and Wu, Xiangqian. (2010) “Deriving an inter-sensor consistent calibration for the AVHRR solar reflectance data record”. International Journal of Remote Sensing, Volume 31, Issue 24, pp.6493-6517.

Hull, J. G., and West. S. G. (1982). “The Discounting Principle in Attribution”, Pers Soc Psychol Bull June 1982 8. pp.208-213.

Karlsson, K.-G. and Dybbroe, A. (2010). “Evaluation of Arctic cloud products from the EUMETSAT Climate Monitoring Satellite Application Facility based on CALIPSO-CALIOP observations”. Atmos.

Chem. Phys., 10, doi:10.5194/acp-10-1789-2010. pp. 1789-1807

Khain, A.P., D. Rosenfeld, and A. Pokrovsky, (2001) “Simulating convective clouds with sustained supercooled liquid water down to –37.5°C using a spectral microphysics model”. Geophys. Res. Lett., 28, 3887–3890.

Kiehl, J., and K. Trenberth. (1997). “Earth’s annual global mean energybudget”. Bull. Am. Meteorol. Soc., 78, pp. 197–208.

Koren, I., Y. J. Kaufman, D. Rosenfeld, L. A. Remer, and Y. Rudich (2005). “Aerosol invigoration and restructuring of Atlantic convective clouds”, Geophys. Res. Lett., 32, L14828,

doi:10.1029/2005GL023187.

Kulmala, M., Vehkamäki, H., Petäjä, T., Dal Maso, M., Lauri, A., Kerminen, V.-M., Birmili, W., and McMurry, PH. (2004). “Formation and growth rates of ultrafine atmospheric particles: A review of observations”. J. Atmos. Sci., 35, pp.143-176.

Le Treut, H., Somerville, R., Cubasch, U., Ding, Y., Mauritzen, C., Mokssit ,A., Peterson, T., and Prather, M., (2007). Historical Overview of Climate Change. in: “Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel

36

on Climate Change”. [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M.

Tignor and H.L. Miller (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. pp.96-110

Leaitch, W.R., Bottenheim, J.W., Bisenthal, T.A., Li, S.-M., Liu, P.S.K., Asalian, K., Dryfhout- Clark, H., Hopper, F. and Brechtel, F. (1999). “A case study of gas-to-particle conversion in an eastern Canadian forest”. J. Geophys. Res., 104, pp.8095-8111.

Lohmann, U. and Feichter J. (2005). “Global indirect aerosol effects: a review”. Atmos. Chem. Phys., 5, 715–737

National Oceanic and Atmospheric Administration (NOAA). (2012). http://www.noaa.gov/

Norris. J.R., and Wild. M. (2007). “Trends in aerosol radiative effects over Europe inferred from observed cloud cover, solar dimming and solar brightening”. J. Geophys. Res., 112. D08214, doi:10.1029/ 2006JD007794.

Penner, J.E. et al. (2001). Aerosols, their Direct and Indirect Effects. “Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change” [Houghton, J.T., et al. (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, pp. 295 – 296.

Ramanathan, V., Krutzen, P. J.,Kiehl, J. T. and Rosenfeld, D. (2001). “Aerosols, climate and hydrological cycle” Science, vol. 294. pp. 2119–2124.

Ramaswamy, V. et al. (2001). Radiative forcing of climate change. “Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change” [Houghton, J.T., et al. (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, pp. 349 – 416.

Roebeling R. A., (2008), “Cloud Properties Retrieval for Climate Studies using SEVIRI and AVHRR data”, PhD Thesis.

Roebeling, R.A., Feijt A.J. and Stammes P. (2006). “Cloud property retrievals for climate monitoring:

implications of differences between SEVIRI on METEOSAT-8 and AVHRR on NOAA-17”. J.

Geophys. Res., 111, D20210, doi:10.1029/2005JD006990.

Roebeling,R.A., H. Hauschildt, D. Jolivet, E. Meijgaard and A.J. Feijt. (2003). “Retrieval and validation of MSG and AVHRR based cloud physical parameters in the CMSAF, Proceedings: published, EUMETSAT data users conference”, Weimar, Germany, EUMETSAT, pp. 183-198

Schulz, J., Albert, P., Behr, H.-D., Caprion, D., Deneke, H., Dewitte, S., Dürr, B., Fuchs, P., Gratzki, A., Hechler, P., Hollmann, R., Johnston, S., Karlsson, K.-G., Manninen, T., Müller, R., Reuter, M., Riihelä, A., Roebeling, R., Selbach, N., Tetzlaff, A., Thomas, W., Werscheck, M., Wolters, E., and Zelenka, A. (2009) “Operational climate monitoring from space: the EUMETSAT Satellite Application Facility on Climate Monitoring (CM-SAF)”. Atmos. Chem. Phys., 9, 1687-1709, doi:10.5194/acp-9-1687-2009

37

STAR (Center for Satellite Application and Research). (2012).

http://www.star.nesdis.noaa.gov/smcd/emb/vci/VH/vh_avhrr_ect.php Swedish Meteorological and Hydrological Institute (SMHI). (2012).

http://www.smhi.se/reflabmodeller/Bakgrund/Luftfororeningar/svaveldioxid-1.19621

Vestreng, V., Myhre, G., Fagerli, H., Reis, S., and Tarrasón, L. (2007). “Tweny-five years of continuous sulphur dioxide emission reduction in Europe”. Atmos. Chem. Phys., 7, 3662-2681.

Voiland, A. (2010). “Aerosols: Tiny Particles, Big Impact”. National Aeronautics and Space Administration (NASA). http://earthobservatory.nasa.gov/Features/Aerosols/page4.php Wild, M. (2009). “Global dimming and brightening: A review”, J. Geophys. Res., vol. 114, no. D16. pp.1- 16

Wild, M., et al. (2005). “From dimming to brightening: Decadal changes in surface solar radiation”. Science, 308. p.847–850, doi:10.1126/science.1103215.

Wild, M., Ohmura A., and Makowski K. (2007), “Impact of global dimming and brightening on global warming”, Geophys. Res. Lett., 34, L04702, doi:10.1029/2006GL028031

Willson, R. C., and Mordvinov A. V. (2003). "Secular total solar irradiance trend during solar cycles 21 – 23”, Geophys. Res. Lett., 30(5), 1199, doi:10.1029/2002GL016038.

38

A

A - L

Abbreviation Meaning Unit

AC All liquid clouds with COT ≥ 3.6 -

AVHRR Advanced Very High Resolution Radiometer -

C1 Liquid clouds with 3.6 ≤ COT ≤ 9.4 -

C2 Liquid clouds with 9.4 < COT ≤ 23 -

C3 Liquid clouds with COT > 23 -

CCN Cloud Condensation Nuclei -

CM-SAF Satellite Application Facility on Climate Monitoring - CNP Liquid clouds with no precipitation and COT ≥ 3.6 -

COT Cloud Optical Thickness Unitless

CTT Cloud Top Temperature Kelvin

CV Coefficient of Variation Unitless

ECMWF European Centre for Medium-Range Weather -

Forecasts

EDGAR European Database for Global Atmospheric Research -

ERB Earth’s Radiation Budget W m^-2

EUMETSAT European Organization for the Exploitation - of Meteorological Satellites

GEBA Global Energy Balance Archive -

GHG Greenhouse Gas -

IN Ice Nuclei -

IPCC Intergovernmental Panel on Climate Change -

NOAA National Oceanic and Atmospheric Administration -

RF Radiative Forcing W m^-2

SSR Surface Solar Radiation W m^-2

STAR Center for Satellite Application and Research -

39

A

B

Appendix B includes a figure with all CV plots.

C3 clouds show a slightly lesser value for the later years (fig. B1:F) than the earlier (fig. B1:C). C2 clouds (fig. B1:B; B1:E) stay approximately unchanged. CV of C1 clouds (fig. B1:A; B1:D) differ much between the periods and show the greatest CV.

Fig. B1 CV of CTT for C1, C2, C3-clouds for 1985-88 and 2004-07.

CV CV CV

CV CV CV