Global dimming : a regional perspective

Camilla Weum Johansen

Institute of Geophysis

University of Oslo

05.05 2005

In spite of inreasing surfae temperature s, a falling trend in surfae insolation has been

observed over the past 50 years at stations worldwide, and a link to anthropogen i aerosols

hasbeen suggested. In the present thesis, reports on this "global dimming" arestudied and

brieyomparedtootherlimatehangessuhasglobalwarmingandatmospheriirulation

hanges. Togetabetterunderstandingoftheonnetionbetweenglobaldimmingandhanges

inloudover, monthly reordsof surfaeinsolationand loudamount ofvaryinglengthsare

investigated forsixstations in north-western Europe.

Theresults indiatethatglobal dimminghasbeen present innorth-western Europeup to

the late 1980s/early 1990s, after whih the dimming has turned to a brightening. Although

the representativity of merely six stations to the tendeny of entire north-western Europe

may be debatable, the results orrespond well to observed hanges in the regional pollution

level: There was heavy industrial build-up (partiularl y in eastern Europe) up to the late

1980s, afterwhih the ollapse ofthe EastBlo and politial interventions to limit pollution

lead to re-improved air quality. Additionally, some stations experiened an inrease in loud

amountspriortoabout1990,sueeded byadereasingtrend, whihwillhaveontributedto

the observed variations in global radiation.

Theorrelation between surfae insolationand loudover wasfoundto be highandsig-

niant at all stations, and did at two of the stations seem to be the dominating ause of

the radiation trend. Itappearsthat insome ases,the eet of aerosolloads andloudover

ollaborate, whileinotherasestheyompete. Asthepresentstudyshows,dimmingmayo-

ur in spite ofimproved aerosollevels,andbrightening mayourinspite ofinreasing loud

overs. The fat thatthe surfae insolationtrend is equallydependent on natural variations

in louds ason human-indued variations in aerosol loads, is an important point as debates

on global dimming in general tend to be fousedonthe eetof anthropogen i pollution.

I wish to thank my external supervisor, assoiate professor Aksel Walløe Hansen (Depart-

ment of Geophysis, University of Copenhagen ), for guidane and inspiration. Manythanks

also to myinternal supervisor, professor Jón Egill Kristjánsson (Department of Geosienes,

University of Oslo), for help and ounelling via e-mail, to Lane Olav Eastgate and Frode

Svendsen for great help with orretions, statistis and language problems, and to Rasmus

Haugaard Nielsen for patient proofreading and moral support.

CamillaWeumJohansen,01/05 2005

Written while guest student at the Universityof Copenhagen,

Niels BohrInstitute Geophysial Department

i

1 Introdution 1

2 Theory 3

2.1 Radiative Transfer . . . 3

2.1.1 General desriptionof radiative energy . . . 3

2.1.2 Absorption and sattering ofshortwave radiation . . . 4

2.1.3 Radiative foring . . . 6

2.1.4 Feedbakmehanisms . . . 7

2.2 Clouds andradiation . . . 8

2.2.1 Eet of louds onsolar radiation. . . 8

2.2.2 Eet of louds oninfrared radiation . . . 8

2.2.3 Cloud hanges andfeedbaks . . . 9

2.3 Aerosols andradiation . . . 10

2.3.1 The aerosoldiret eet . . . 11

2.3.2 The aerosolindiret eet . . . 11

2.3.3 Aerosolhanges andfeedbaks . . . 12

3 Climate hanges and global dimming 14 3.1 Earth's energy balane . . . 14

3.1.1 The natural greenhouseeet . . . 15

3.2 Naturallimate hanges . . . 15

3.2.1 External auses oflimate hange . . . 16

3.2.2 Internal auses oflimate hange . . . 17

3.3 Anthropogeni limate hanges . . . 18

3.4 Reenthanges in numbersand gures . . . 19

3.5 The roleof global dimming . . . 24

3.5.1 The magnitudeof the dimming . . . 24

3.5.2 Geographial distribution . . . 25

3.5.3 Possible auses . . . 26

3.5.4 The global dimming/warming paradox . . . 29

3.5.5 Impliations of global dimming . . . 30

4 Measurements, Stations and Data analysis 33 4.1 Instrumentsand observation methods . . . 33

4.1.1 Measuring temperature . . . 33

4.1.2 Measuring radiation . . . 34

4.1.3 Measuring loudamount . . . 34

4.2 The stations . . . 35

4.2.1 Hamburg station . . . 35

4.2.2 Copenhagenstation . . . 36

4.2.3 Lund station . . . 37

4.2.6 Luleå station . . . 38

4.3 Statistial Analysis ofthe data . . . 38

4.3.1 Correlation . . . 38

4.3.2 Signiane of orrelations. . . 39

4.3.3 Finding trends . . . 39

4.3.4 Autoorrelation . . . 40

4.3.5 The importane of orretinterpretation . . . 41

5 Results 43 5.1 Data fromHamburg . . . 44

5.1.1 Correlation between loudamount andglobal radiation in Hamburg . . 46

5.2 Data fromCopenhagen. . . 47

5.2.1 Correlation between loudamount andglobal radiation in Copenhagen . 49 5.3 Data fromLund . . . 50

5.3.1 Correlation between loudamount andglobal radiation in Lund. . . 52

5.4 Data fromAberdeen . . . 52

5.4.1 Correlation between loudamount andglobal radiation in Aberdeen . . 54

5.5 Data fromBergen. . . 55

5.5.1 Correlation between loudamount andglobal radiation in Bergen . . . . 56

5.6 Data fromLuleå . . . 57

5.6.1 Correlation between loudamount andglobal radiation in Luleå . . . . 58

6 Disussion 60 6.1 The signiane ofloud amountsfor global dimming. . . 60

6.2 Comment on data quality . . . 64

6.3 Comparing the 1983 to 2003 trends at the sixstations . . . 66

6.3.1 The Copenhagen -Lund disrepany . . . 66

6.4 Representativityofthese sixstations fornorth-western Europe dimming . . . . 67

6.5 The auseof global dimminga reevaluation . . . 68

7 Summary and Conlusion 70 7.1 Futureperspetive . . . 72

A Tables of statistial values 74 A.1 Correlation oeients forglobal radiation and loudamount . . . 74

A.2 1983 to 2003 global radiation trends . . . 75

A.3 1983 to 2003 loud amount trends. . . 75

A.4 Monthly means ofloud amount . . . 76

A.5 Monthly preipitatio n hanges . . . 76

B Supplementary plots 77 B.1 1983 to 2003 global radiation for allstations . . . 77

B.2 1983 to 2003 loud amountsfor all stations . . . 78

B.3 ISCCP plots . . . 78

Bibliography 79

Introdution

Partiularly sine the industrial revolution in the mid 18th entury, human ativities have

had an inreasing eet on limate, most importantly by adding to the onentration of

numerousatmospherionstituents. Possiblyasaresultofthis,the Earthhasexperienedan

inrease in global surfaetemperatures (0.6

◦

Csine the late 19th entury[Intergovernmental

Panel on Climate Change (IPCC), 2001, setion 2.3℄) and oean temperatures (0.15

◦

C from

1958 to 1998 [Levitus et al., 2000℄). But while sientists generally agree that the limate is

warming,theydisagreeontheauses,ontheexpetedonsequenes,andevenonobservations

oftheseonsequenes. Suhnononformitiesanusuallyberelatedtodierenesinmeasuring

instruments, methods of observation or struture of models. Although a quite natural and

inevitable part of researh, this an often turn sienti onlusions into a question of mere

subjetive assessment.

Examples of diverging observations of limate hanges are many. Some argue that the

warming is followed by inreased evaporation from land and water surfaes [Brutsaert and

Parlange,1998;Golubevetal.,2001℄,strengthening thehydrologialyleandinreasingpre-

ipitation. Otherslaimtohaveprooffor adereasein theterrestrial evaporationomponent

[Petersonetal.,1995;RoderikandFarquhar,2002;Liepertetal.,2004℄, whihonverselywill

Figure 1.1: A radiant sun above

the Larsbree n glaier, Svalbard.

Photo: CamillaW.Johansen.

be assoiated with a spin-down of the hydrologial yle and

dereased preipitatio n. While Pallé et al. [2004℄ speak of a

dereasein Earth'sreetivitydue todereased loudalbedo,

Liepertet al.[2004℄arguethatloudoptialdepth(andhene

loud albedo)isinreasing.

Amongst all these oniting observations, sientists have

beome aware of what seems to be a global derease in the

amountofsolarradiation reahingthesurfae. Thisdisovery

is intriguing espeially in onnetion with the observations of

globalwarming,beauseitontraditstheommonpereption

thatsurfaetemperature sandsurfaeinsolationarepositively

orrelated. Thephenomenonwasaddressedasearlyasin1974

bytheIsraelisientistSuraqui[Suraquietal.,1974℄, andagain

in1989byAtsumuOhmura,sientistattheSwissUniversityof

Tehnology. Butalthoughbasedonatual observations, these

publiationsweremetwith skeptiismamongothersientists,

who argued that sine temperature s and (aording to some)

evaporation were inreasing, the obviously ontraditor y and

surprisingly large fall in solar radiation simply ould not be

real. Consequently, the phenomenon was widely regarded as a result of erroneous measure-

ments or not regarded at all, and the subjet remained a matter of ontroversy for the next

deade. Slowly,however,thedisoveriesofSuraquiandOhmura weresupportedbyanumber

of studies worldwide [Gilgen et al., 1998; Stanhill and Cohen, 2001; Liepert, 2002℄, and in

2002, Roderik and Farquhar [2002℄ onneted the observations of dereasing pan evapora-

tion to the dereasing radiation. Now being linked with other physial proesses and more

extensiveobservations,thedereasingtrendinsurfaeinsolationwasbeomingmoreoherent

andredible. Thephenomenon wasnallyaknowledged asgenuine, and wasdubbed Global

Dimming .

Themagnitude ofthe global dimming is generallyonthe orderof afew perent: Stanhill

and Cohen [2001℄ set the globally averaged redution to 13.5% (20

W m ⁻²

years,Liepert[2002℄estimatesa4%(7

W m ⁻²

et al. [1998℄ onlude that the global mean hange adds up to 6% (10.2

W m ⁻²

to 1990. Thereview of Stanhilland Cohen [2001℄ whih summarizes the results ofa number

of studieson global dimming,onludesthatthe most probable auseofthe dimming isthat

"inreasesin man madeaerosols and other pollutantshave hanged the optial properties of

the atmosphere,in partiularthoseoflouds". Although aplausibleexplanation andawidely

aepted one, an earlier study by the same authors [Stanhill and Cohen, 1997℄ desribes a

signiant solar radiation derease at Antartia, where in the same period loud over and

sea ie did not inrease and there was little evidene for inrease in aerosols. The report

oers no other possible explanation and leaves the ause of the dimming an unanswered

enigma. Liepert [2002℄ attributes only a small fration of the dimming to anthropogen i

aerosols, asribing the rest to a possible greenhouse gas warming-rel ated inrease in loud

optial depth. Clearly, there is not a omplete onordane asto what auses the dimming,

and there arestill septiswho attributegreat parts ofit toinstrument errors.

Today, global dimming is reeiving more attention from the sienti ommunity and is

even starting to attrat the interest of media and the general publi. But we are only just

beginning to get a lear idea of its extent, of its role in limate hange and of its auses.

If the trend is, in fat, a global phenomenon , it must play an important role in the global

energy budget, andit maybeinteronneted with global warming in ways yet unknown. An

exploration ofthe trendin surfaesolar radiationand itsrelation totemperature, louds and

aerosols mayshednewlighton howthe omponentsof thelimate systeminterat,mayhelp

distinguish the anthropogen i ontribution s to limate warming from ontribution s owing to

natural variations, andmay improveour abilities to predit futurelimate hanges.

In this study, the surfae trends of shortwave radiation for various time intervals will be

studied at six dierent stations in north-western Europe. The trends and variations of the

surfae radiationwill thenbeompared to thetrends andvariations oftotalloudover,and

for some stations also preipitatio n. Additionally, onentration or emission levelsof sulphur

dioxide arepresented forall loationsto getan impression ofaerosol onentrations. Results

show a lear tendeny for dereasing radiation up to around the late-1980s, after whih an

inreasing trendis seen at all but one station. The eet of hanging loud amounts does at

some loations seem the dominating ause of the variations in surfae insolation. At other

loations,thebakgroundsignalofpollutionandaerosolsappearstobetheontrollingfator.

Thefollowinghapterprovidessomebasiaspetsofradiativetransfer,inludingadesrip-

tion of the radiative properties of louds and aerosols. Chapter 3 will give a more thorough

introdution to limate hanges and the role of global dimming in relation to the hanges

observed. A review of methods for measurements, statistial analysis, and a desription of

the stations will be given in Chapter 4, while the results will be presented in Chapter 5. In

Chapter 6, the results are disussed in onnetion with the bakground material, and nally

a onlusion isdrawnin Chapter 7.

Theory

2.1 Radiative Transfer

I willin this setionpresenta briefoverviewofthe fundamentalpriniplesof radiativetrans-

fer, whih serves as a mehanism for energy exhange within the earth-atmosphere system

and between it and the rest of the universe. The amount of energy that the system reeives

from the Sun and ultimately the temperature at Earth's surfae, depends on the radiative

properties of all the omponents of the limate system. Below follows a short desription of

radiative energy and of the transmittane of solar radiation through the atmosphere, based

on the theory of Hartmann [1994, Chapter 3℄. The priniples of radiative foring and feed-

bak mehanisms will then be dened, after whihI in separate setions will aount for the

radiative properties oflouds and aerosols.

2.1.1 General desription of radiative energy

Theenergyofradiationmaybeexpressedintermsofitsradiane orintensity. Monohromati

intensity, whih desribes the amount of radiant energy (d

F _λ

(

λ

λ + dλ

dA

dω

diretionin a time interval

dt

dF _λ = I _λ cosθdωdAdλdt,

Here,

I _λ

λ

θ

asthe angle between the radiation and the normal to the surfae, and

dω

denedby

dω = sinθdθdφ

obtain suhaquantity, weusethe denitionof thesolidangleintoequation2.1andintegrate

over the upper hemisphere. The resulting

F _λ

F _λ = Z 2 π

0

Z ^π ₂

0

I _λ (θ, φ)cosθsinθdθdφ

Ifdesirable,we an also integrate over allwavelengths to obtainthe total uxdensity

F

F =

Z ∞

0

F _λ dλ

Theuxdensityhasunitsofwattspermetersquared(

W m ⁻²

of energy (of all wavelengths) that passes through a plane surfae of unit area per seond.

1

1

Aswillbepratiedintheurrentstudy,thisunitanalsobetransformedtoJoule ,whihisameasureof

thetotalenergyreeivedoveragivenperiodoftime.

M J m ^{− 2}

M J

10 ⁶ J

Figure2.1: An illustration showing the angles that dene the radiane owing through a unit

area

dA

θ

φ

dω

Theux densityfrom the Sun that reahes the top of Earth's atmosphere is alled the solar

onstant, andhasavalueof approximately 1367

W m ⁻²

radiationis subjet to asittravels troughthe Earth'satmosphere dependson the reeting,

absorbingandsatteringpropertiesoftheatmospherionstituentsaswillbedisussedbelow.

2.1.2 Absorption and sattering of shortwave radiation

How is inoming shortwave radiation from the Sun depleted as it travels through the at-

mosphere? Thedegreeofdepletionisdependent onthe lengthofthetraveledpathandonthe

absorbing and sattering qualities of the gases, louds and aerosols within that path. These

qualities of amedium anbedesribedbythe extintionoeient

β _ext

the absorption oeientand the sattering oeient:

β _ext = β _abs + β _sct

a beam of radiation passing through a layer of depth

dz

dz = z 2 − z 1

onsider the optial depth

τ

dz

Notethatthislayermayverywellomprise theentiredepth ofthe atmosphere,in whihase

z 1 = z _{surf ace} = 0

z 2 = z ∞ = z _{T OA}

T OA

τ = Z z 2

z 1

β _ext dz dτ = − β _ext dz

Now, onsider the attenuation of radiation over an innitesimal path

ds

enough sothat:

-the extintion oeient

β _ext

-the inident radiation isattenuatedbyaninnitesimal amount

dF _λ

Thehange in radiationis then givenby:

dF _λ = F _λ (s + ds) − F _λ (s)

where

F _λ (s)

s

z 2

F _λ (s + ds)

s + ds

Figure2.2: Illustrationoftheextintionpathofasolarradiationbeamthroughaplane-parallel

atmosphere (from Hartmann[1994, p.53℄).

or altitude

z 1

radiation at startingpoint

s

us the following expression:

dF _λ = − β _ext F _λ (s)ds

As the atmosphere ishighly stratied, the distribution of gases, louds and aerosols is muh

more variable in the vertial than in the horizontal. Thisallows us for radiative purposes to

treat theatmosphere asplane parallel;at agivenloation weignore thehorizontalvariability

in the struture of the atmosphere,and assume insteadthat all relevant radiative properties

depend stritly on the vertial oordinate

z

the Earth's urvature omes into play, another simpliation may be made by ignoring the

spheriityof the Earth. This is a fair approximation for radiation transporting energy more

or lessvertially throughthe atmosphere. As seenfrom Figure 2.2, the use of the geometri

relationship

dz = − cosθds

dF _λ = β _ext F _λ dz cosθ

Making use of the denition of the optial depth

τ

spetral ux,we an nowwrite:

dF = − dτ cosθ F 1

F dF = − 1 cosθ dτ

Integrating fromlevel

z 2

z 1

Z _z 1

z 2

1

F dF = − 1 cosθ

Z _z 1

z 2

dτ

ln( F _z 1

F _z 2

) = − 1

cosθ τ _z

This, nally, gives us an equation desribing the deay of a beam of shortwave radiation

propagating downwards throughthe atmosphere between altitude

z 2

z 1

F _z 1 = F _z 2 e ⁻

τz1 −τz2

| cosθ |

Using instead the

T OA

z = ∞

z 2

z

z 1

extintion or Beer's law:

F _z = F ∞ e ⁻

τz

| cosθ |

(2.3)

As evident fromthe equation, the radiant uxdeays exponentiallydepending onthe zenith

angle and the optial depth of the layer. In the present study we will be partiularly inter-

ested in the part of shortwave radiation that reahes the surfae, ommonly known asglobal

radiation. Global radiation is the uxof both diret and diusesolar radiation reahingthe

Earth's surfae, and will also be referred to as surfae insolation or surfae solar radiation.

In equation 2.3, global radiation an be obtained by letting

z

from the equation, a redution in global radiation, whih is the objet of the present study,

an be obtained either by a derease in

F ∞

τ

F ∞

in the radiation emittedfrom the Sun, while the atmospheri transmittane or optialdepth

τ

2.1.3 Radiative foring

In an equilibrium limate state, the average net radiation at the top of the atmosphere is

zero and there is a balane between the radiation oming into the system and the radiation

goingout. A hange inoutgoing orinomingradiationwillalter thenet radiation,ausing an

imbalanewhihisreferredtoasradiativeforing [IPCC,2001,Setion1.2.1℄. Inpratie,the

top of the troposphere (the tropopause) is onsidered as the top of the atmosphere, beause

the stratosphere adjusts relatively fast (in a matter of months) to hanges in the radiative

balane, whereas the surfae-troposphere system adjusts muh more slowly, prinipally due

to the large thermal inertia of the oeans. The radiative foring of the surfae-troposphere

system is then dened as the hange in net irradiane at the tropopause after allowing for

stratospheri temperature sto re-adjust to radiativeequilibrium [IPCC, 2001,Setion 1.2.1℄.

Theradiative foring anbeeither positiveor negative, depending onwhetherthe imbal-

ane ausesanexessordeitofenergy withinthelimatesystem,respetively. Thelimate

system responds to radiative foring so as to re-establish the energy balane, inreasing or

dereasing thetemperaturein ordertoinreaseor dereasetheamount ofoutgoingradiation.

This is why a positive radiative foring tends on average to warm the surfae of the Earth

while anegative foringtendsonaverageto oolthe surfae. Changesinsurfae limateisin

other words drivenbyhangesin the balane at the top ofthe atmosphere.

Foringfromagivengas,partileorotheronstituentofthe limatesystem(heredenoted

omponent forsimpliity)atagivenwavelengthregionisgivenin

W m ⁻²

foringinordereithertoexplaintheeetofthe preseneofagivenomponent,ortoexplain

the eet ofhanges in thatomponent. Themagnitudeofthe foringin these twoases an

be understood respetivelyas...

•

...theontributiontotheurrenttop-of-atmosphere(

T OA

more or less energy the system ontains ompared to what it would in the absene of

thisomponent. Forexample,whenwesaythatthetotalradiativeforingduetolouds

is about -17

W m ⁻²

in imbalane at the

T OA

louds depriveus of. Had therebeen nolouds,theEarthwouldbereeiving17

W m ⁻²

more energy, whih would inrease the surfae temperature and hene the amount of

outgoing radiation. Thenet radiationat

T OA

be warmer.

•

...thenetradiativeresultofahangeintheonentrationorpropertiesofagivenompo-

nent over aperiod in time. For example,due to industrialativitythe onentration of

CO 2

auseapositiveradiativeforingof1.46

W m ⁻²

there is in fat a positive imbalane at the

T OA

of surfae temperatures show, the limate system hasalready started to adjust to this

imbalane by warming. If we managed to stabilize greenhouse gas emissions, the net

radiation at the

T OA

out andremainat anewequilibriumtemperature. Thenew equilibriumstatewouldbe

warmer than the pre-industrialone,but would nolonger beradiatively fored.

AnyhangesintheradiativebalaneoftheEarth,inludingthoseduetoaninreaseingreen-

housegasesor in aerosols,will to someextent alter the surfaetemperature. Thisis likely to

aetthehydrologialyleandatmospheriandoeaniirulations,therebyaltering global

andregionalweatherpatterns. Thesieneofalulating andprediting radiativeforingand

feedbaksis thereforean imperativepart ofunderstanding limate hanges.

2.1.4 Feedbak mehanisms

Asmentionedabove,the limatesystemrespondsto aradiative foring byadjustingitstem-

peraturereduingtheimbalanebydoingso. However,atthesametimeasthetemperature

hangesinresponsetoagivenradiativeforing,otherqualitiesdependentontemperaturemay

also hange, whih further alters the net emission of radiation to spae. A positive feedbak

supports the initial foring, allowing the limate system to reah the new equilibrium tem-

perature faster, while a negative feedbak weakens the initial foring and ounterats the

temperature hange. An example of a positive feedbak mehanism is the redution in sea

ieand snowover as the limate warms, whihwill lower the Earth'sreetivityallowing it

to absorb more radiation and inreasing its temperature further. An example of a negative

feedbakmehanism isan inrease in loud over that ould result indiretly from inreased

temperatures dueto enhanedatmospherimoisture levels (itisnotlear whetherornot this

really is the ase). The inreased loud over will inrease the amount of reeted radiation

tospaeandattenuatetheinitial temperature inrease 2

. Duetothisinterlinked natureofthe

limate system, itan beextremely diult to assoiatea partiular observed foring with a

partiular ause. In order to predit foringsdue to hangesin spei limate omponents,

one must therefore know both how this omponent interats with radiation and also all the

feedbakmehanisms itis assoiated with.

2

Aswillbeseeninthefollowingsetion,theeetofhangingloudamountsontemperatureisinreality

2.2 Clouds and radiation

Aording to Hartmann[1994, p.249℄, halfof the Earth's surfae is on average overshad-

owed by louds. As louds to a great

extent aet the amount of energy en-

tering andleaving the Earth-atmosphere

system,itisimportanttogainknowledge

oflouds'rolein theenergybalane. De-

pendingontheirmirophysialproperties

suh as loud water ontent (for liquid

water as well as for ie), loud droplet

(orrystal) radius or shape, their distri-

butionin spaeandthe albedooftheun-

derlying surfae, louds aet radiation

in more than one way. Clouds an be

eient reetors of solar radiation, but

analsotrapoutgoinginfraredradiation.

Figure2.3: Multilayeredloudssurroundingabeautifulu-

mulonimbusanvilsystem. Photo: CamillaW.Johansen.

The net eet of louds on surfae and atmospheri temperatures are deided by how these

two proesses are balaned. Below follows a survey on how louds aet solar and infrared

radiation, respetively.

2.2.1 Eet of louds on solar radiation

Howwelllouds reet solarradiation depends onthe loud'salbedo or reetivity, whih

has a range of 30 to 50% for thin louds and 60 to 90% for thik louds. Figure 2.4

Figure 2.4: The dependene of loud albedo on

loud liquid water path and solar zenith angle

(from Hartmann[1994,p.65℄).

shows the onnetion between liquid water path

(

LW P

of water present in an air olumn ofgiven surfae

area,andalbedo,dependingontheangleofthein-

oming solar radiation: Themoredensethe loud

(and thelargerthe solarzenithangle),thegreater

the albedoandhenethe amountofreeted sun-

light. Although variationsin the loudalbedo are

dominated byvariations in the olumn amount of

liquid water and ie in the loud, the albedo is

alsosensitiveto thedroplet size[Hartmann,1994,

p.65℄. The albedo is greatest for small droplets,

prinipally beausefor a givenmassthese over a

larger surfae area. Clouds will also (espeially if

they are deep) satter radiation in all diretions,

a proess that will throw some radiation bak to

spae, some downtowardthe surfae andsome to

be absorbed bygases or other loudsin the atmosphere.

As louds reet solar radiation bak to spae, the limate system is deprived of energy

that would have ated to warm it. In terms of top-of-atmosphere radiative foring, IPCC

[2001, Setion 14.2.3.1℄ onlude thatthis albedo eet hasamagnitude ofabout -50

W m ⁻²

2.2.2 Eet of louds on infrared radiation

Clouds also have the well known property of "trapping" outgoing infrared radiation by ab-

sorbing andre-emittingit, theeienyofwhihalsodependsonthe

LW P

Average Cloud-free Cloud foring

Outgoing infrared radiation [

W m ⁻²

Absorbed solar radiation [

W m ⁻²

Net radiation [

W m ⁻²

Albedo [%℄ 30 15 +15

Table2.1: Cloud radiative foring as estimated from satellite measurements (from Hartmann

[1994,p.75℄).

about 30

W m ⁻²

radiation that isre-emitted baktoward the surfae. But due to the atmospheri lapse rate

wheretemperature dereaseswith altitude, warming also ours whenradiation isre-emitted

in the opposite diretion; into spae: Aording to Stefan-Boltzmann's law that will be de-

ned on page 15, radiation emitted from bodies with low temperature ontains less energy

than radiation emitted from warmer bodies, and as the louds are older than the Earth's

surfae, less energy is emitted to spae in a loudy than in alear-sky situation. Losing less

energy, the atmosphere isthus warmed.

AsshownbyFigure2.5,loudsbeomeopaque toinfrared radiation(thatis, noradiation

Figure 2.5: The dependene of the long wave

emissivity on loud liquid water ontent (from

Hartmann[1994,p.66℄).

is transmittedor reeted; all isabsorbedand re-

emitted) at

LW P

gm ⁻²

louds (e.g., a layer of solid stratoumulus) are

thereforethemosteientabsorbersandemitthe

most energy bakto the surfae, but high louds

(e.g., irrus) are the oldest and thus the best at

onservingtheenergy withingthe limatesystem.

However, the eet of a loud systemon infrared

radiation isoftenambiguous beause,like the one

in Figure 2.3 on the preeding page, they often

onsist of multiple types of louds. The net ef-

fet of louds on the radiation budget is ompli-

ated further by the fat that louds aet solar

and infrared radiation simultaneously, and more-

over by the fat that the eets louds have on

temperature, may aet the nature of the louds

themselvesin positiveand negative feedbaks.

2.2.3 Cloud hanges and feedbaks

It is generally aepted that the net inuene of the two opposing eets of louds on the

radiation balane is negative; the albedo eet of louds exeeds the eet of long wave

absorption. Aording to Hartmann[1994℄,the valueofthe net foringliessomewhere inthe

proximity of-20

W m ⁻²

But while the overall net eet of louds on the radiation balane is known, the sign

and magnitude of loud feedbaks with respet to global warming, inreased pollution and

other limate hanges is not. Many people assume that an inreased evaporation following

from the limate warming 3

, and the resulting inrease of moisture in the atmosphere, ould

inrease the amount of louds [Aguado and Burt, 1999℄, whih would have a ooling eet.

However,an airparel in awarmer atmosphere needsmore watervapormoleulesto beome

3

Whetherornotevaporationfromland willinreaseasaresultofinreased globaltemperatureis aon-

troversialissue,aswillbedisussedindetailinChapter3. Evaporationfromthewarmingoeansarehowever

saturated and ondensate into louds, whih makes it hard to antiipate the exat eet of

inreasedevaporationonlouds. AordingtoIPCC[2001,Setion7.2.2.4℄,shemesprediting

loudinessasafuntionofrelativehumiditygenerallyshowanupwarddisplaementofthehigh

loudover inresponsetoagreenhousewarming,resultingin apositive feedbak. Analysesof

satellite data reveal aninrease in both lowand irrusloud optialdepth with temperature

[TselioudisandRossow,1994℄. Asinreasesinlowandirrusoptialdepthinduefeedbaksof

oppositesigns,the reportonludesthatthe nethange inawarmerlimate isunertainand

relieson other thermodynami and dynamionditions. The dependene ofloud properties

on temperature was also demonstrated in a model simulation by Liepert et al. [2004℄, who

foundinreasedtotalloudoptialdepthssinepre-industrialtimes,assoiatedwithenhaned

availability of atmospheri moisture. Studies have also been made on the eet of pollution

of louds: Kaufmann and Freedman [1999℄ found thatlouds nearurban areas where smoke

partileswereabundant,haddereaseddroplet sizeandheneinreasedreetivityompared

to louds in rural environments. Meanwhile, Rosenfeld [2000℄ showed that pollution also

suppresses preipitatio n (both in form asrainand snow) in louds, a proess whih prolongs

the loud'slifetime. Althoughexperimentsandmodelsimulations arearehelping to mapthe

eet of limate hangeson louds,the piture is yet far fromlear.

How, then, will the limate respond to hanges in loudiness? An estimatebyHartmann

[1994, p.249℄ onludesthat the eet of aninrease in total loudiness by10%(for example

from50%meanloudoverageto60%)willhavethe samemagnitudeasadoublingofarbon

dioxide onentration, osettingthe eet of

CO 2

rudeomparison betweennumbers, anddoesnotregardneither loudpropertiesnordistrib-

ution. Aswehaveseenfromabove,dierenttypesofloudsaetradiation dierently, anda

limate responsewill depend on whattypesof loudshange and whetherthey beome more

or lessabundant,thiker or thinnerand higher orlower inaltitude. Forexample,an inrease

in high loudiness will have a net warming eet, but a simultaneous inrease in high loud

density (and hene albedo) will perhaps outweigh this warming. The distribution of louds

above reetive and not soreetive surfaesalsoplaysa role.

Needless to say, the issue of limati eets of loud hanges is highly omplex and the

ability to alulate the eet of hanges in loudiness requires rst of all a better loud ob-

servation systemin whih louds in all layers and their propertiesan be reorded with high

time-and spae-resolution. Suhasystemisurrentlynot available, andwithoutknowing at

leastthe atualfrationsofhigh,middleandlowloudover,keepingtrakofanyhanges in

the above beomesa diult task.

2.3 Aerosols and radiation

Aording to Wallae and Hobbs [1990, p.114℄, an aerosol is dened as "a suspen-

Figure2.6: Thestableboundarylayerhastrappedthean-

thropoge niaerosolswhihlayasabrownishhazeabovethe

ityof Oslo. Photo: CamillaW.Johansen.

sion ofsolid or liquid matter in agaseous

medium" (in our ase air). The partiles

rangeinsizefromabout

10 ⁻⁴ µm

tens of mirometer s, and are formed ei-

therbydispersalofmaterialatthe surfae

or byhemial transformations of preur-

sorgases [Hartmann, 1994,p.291℄. Wind-

blown dust, sea salt and gas-to-partile

onversions are the most ommon nat-

ural soures of aerosols, whih are on-

stantly supplied to and removed fromthe

throughindustry, biomass burningand tra (seeFigure 2.6onthe preeding page).

Although theirinuene on theatmospheri energy transferissmallerthan thatof louds

andgases(seeTable2.2),aerosolsplayanimportantrolebysattering andabsorbingsolarra-

diation andbyalteringthe lifetimeandpropertiesoflouds. Howaerosolsaettheradiation

budgetdependsonpartilesizeandomposition,onspatialdistribution, andontheeieny

of potential aerosol-loud interations. These proesses areomplex and the fatorsinvolved

arehighlyvariableinspaeandtime. Asaresult,therearelargeunertaintie sonnetedwith

estimations of the magnitude of aerosol foring, and these unertainties are propagated into

preditions oflimate feedbakto atmospheriaerosol hanges. Thefollowing pageswill give

a presentation of the role of aerosols in the radiation budgetand briey disuss estimates of

foring and limate response.

2.3.1 The aerosol diret eet

Changesintheradiationbudgetarisingfromsatteringandabsorptionbyaerosolsarereferred

to as the diret radiative eet or diret foring of aerosols. Depending on partile size and

hemial omposition, atmospheri partiles have dierent abilities to inuene solar and

infrared radiation, and the sign of the foring depends on how eient a given aerosol is in

sattering and/or absorbingradiation. As most aerosols do both, itis useful to onsiderthe

single sattering albedo (

SSA

absorption. Blakarbonaerosolsdoatvisiblewavelengthshave

SSA

thata large partof their interation with radiation omes from absorption. Sulfate aerosols,

on the other hand, have

SSA

to Ramanathan et al. [2001℄, aerosols have a net negative top of atmosphere (

T OA

when the

SSA

T OA

SSA

SSA

signof the net eet depends onother fatorslike loudfration, surfae albedo and vertial

aerosoland loud distribution. Asa onsequene,

T OA

publiations are highly unertain. Estimates are however made, and IPCC [2001, Setion

6.7.1℄haveonsidered individualaerosol speies,ndingthe diret aerosol foring tobe +0.1

W m ⁻²

W m ⁻²

W m ²

for sulphate aerosols. Similar results arefound by Harvey[2000, p.108℄ and Bouher [1995℄,

illustrating that, as most sientists agree on, the net diret eet of all aerosols is a slight

ooling.

2.3.2 The aerosol indiret eet

Aerosols may also aet limate indiretly through their interation with louds in the at-

mosphere. An important role is played by aerosols in loud ondensation, where they fun-

tionasloudondensation nulei (

CCN

0.05µm

CCN

Solar radiationTerrestrial radiation

Atmospheri onstituent Absorption Sattering Absorption/emission Sattering

Air moleules 1 2 1 3

Aerosols 2 2 2 3

Clouds 2 1 1 3

Table 2.2: Relative importane of various radiative transfer proesses in the global energy

balane (fromWallae and Hobbs [1990, p.308℄): 1 denotes proesses of primaryimportane,

Soure type Natural soures Anthropogeni soures

Diret emission 1540

T g year ⁻¹

T g year ⁻¹

Gas-to-partile onversion 394

T g year ⁻¹

T g year ⁻¹

Total 1964

T g year ⁻¹

T g year ⁻¹

Table 2.3: Soures of aerosols by mass, fromHartmann[1994, Table 11.1℄.

alsoontribute[Ramanathanetal.,2001℄. Aninreasein

CCN

lead to an inrease in the number of loud drops (and ie rystals) in louds [IPCC, 2001,

Setion5.3.2℄. Thistendstoinreasethe loudalbedoasexplainedinSetion2.2.1onpage8,

resultingin aooling ofthe surfaethroughinreasedreetionofsolar radiation. Thiseet

isreferredto asthe rst indireteet .

Assumingthatthe ondensedloudmoisture staysonstant,an inreaseinthe number of

dropletswillbeassoiated with adereasein louddroplet radius. A dereasein droplet size

willinhibitthe loud'spreipitatio neieny, therebyprolongingtheloudlifetimeresulting

in inreased amounts of reeted solar radiation [Rosenfeld, 2000℄. This (ooling) eet is

alled the seond indiret eet .

Yet another eet arises from solar heating of the boundary layer by blak arbon ab-

sorption,whih anevaporatesome ofthe louds andallowmore solarradiation to reahthe

surfae. Thispositive surfae foring isalled the semidiret eet of louds.

Harvey [2000, p.109℄ estimates the indiret eet due to sulphate aerosols to lie in the

interval -0.5 to -2.2

W m ⁻²

of aerosolis ooling, andobservations even showthat the

T OA

indireteetissigniantlylargerthanthe

T OA

foring at the surfae omes from the diret eet. Little fous has been direted to the

foringsofthe seond indireteet and the semidireteet of loudsproesses thatmay

verywell ontributesigniantly to the aerosoleet on limate.

Natural versus anthropogeni aerosol loading

Thetotal ux of natural aerosols averaged over the globe isabout three to four timeslarger

thanthe uxofaerosols generatedbyhumanativities[Hartmann,1994, p.293℄. Thisanbe

seen in Table 2.3. Sine aerosol lifetimes are relatively short (days to weeks [Harvey, 2000,

p.33℄), the geographial distribution varies substantially, leaving onentrations partiularly

highlose to soures. Asa resultof this,the uxofanthropogen i aerosols an ona regional

sale be upto ve timeslarger than the natural aerosol ux[Srinivasan andGadgil, 2002℄.

Thelargest ontribution to humanindued aerosolloading omes from sulphateaerosols.

Sulphate aerosols are madein gas-to-partileonversionsfrom sulphur gases (

SO 2

emitted throughtherening andombustion ofsulphur-ontain ing oalandfromthe melting

of sulphur-mixed minerals suh as zin, opper and lead [Harvey, 2000, p.41℄. Estimates of

hangesin anthropogen iaerosolemissions sinethe industrialrevolutionwill be given inthe

following hapter.

2.3.3 Aerosol hanges and feedbaks

Altogether,aerosolsexertaooling eeton limate. Asillustratedin Figure2.7onthe next

page,thesurfaeoolingduetoaerosolsisexeededbythesurfaewarmingduetogreenhouse

gases (see

T OA

loadmayalterthis relationship,thus aetingthedegreeofwarming. Butasalways,hanges

in aerosolonentration, propertiesor distribution do not only aet limate diretly but do

also eliit feedbakmehanisms. Ramanathan et al. [2001℄ mentions three important aerosol

feedbaksrelated to the seondindiret eet:

1. As preipitatio n is the main soure of aerosol removal [Harvey, 2000, p.33℄, the sup-

pression of preipitatio n by aerosols through the seond indiret eet prolongs the

atmospheri lifetime of the partilesand henethe magnitude of their limati impat

(positive feedbak).

2. The derease in preipitatio n will lead to drier surfae onditions, whih will produe

higher amounts ofwindblown dustand also inrease the number of forest res, adding

further to the aerosolonentration (positive feedbak).

3. Finally,suppressionofpreipitatio nindeeponvetiveloudsespeiallyinthetropis

allows for transport of more water and aerosols into the upper troposphere and lower

stratosphere, inreasing the watervapor greenhouseeet in this altitude and possibly

osettingsome of the aerosolooling (negative feedbak).

Similar eetsan of ourse alsotake plaewithout an initial aerosolhange: Awarmer and

more moist global limate will byinferene dampen at least the natural aerosol soures, and

shorten aerosollifetimes. But aswill be seenin the following hapter, hanges in irulation

patterns aet preipitatio n levels dierently from region to region. Ultimately, the eet of

limate hangesonaerosols willthereforebe determinedbywhatkindsofhangesournear

the major emissionsoures.

As the Earth's population is growing, one an imagine a following inrease in biomass

burning, useof fossilfuels and deforesting,all of whih willadd to the amount of aerosolsin

the atmosphere. Our urrent understanding of the aerosol-loud-radiat ion proesses are by

far insuient, andmore researh isneededto nd waysto more auratelyestimateaerosol

foring and feedbaks.

Figure2.7: Comparison of global annual mean anthropogeni aerosol foring and greenhouse

foring (from Ramanathan et al. [2001℄).

Climate hanges and global dimming

Climate is to some extent always hanging. Variations in the Earth's path around the Sun,

volanieruptions,hangesinatmospheriomposition,oeanirulationhangesandvarying

energy output from the Sun are only some of the auses of the natural variations in limate

thatthe Earth has experiened ever sine its reation roughly4.5 billion years ago. Human

beings have only been a partof this for the last few hundred thousand years,yet asmodern

limate may be demonstrating, we are already beginning to ause limate hanges that are

omparablewith the natural ones.

As mankind's ativities are spreading to even the most remote plaes (see Figure

Figure3.1: SootertraksontrastingthewildandunspoiltBlaahuken

mountain ontheislandof Svalbard. Photo: CamillaW.Johansen.

3.1), we are indubitably start-

ing to set our mark on Earth's

limate. This makes it inreas-

ingly important to fully under-

standthelimatesystemandits

omponents. Globaldimmingis

one of the many proesses that

arenotyet wellunderstood, but

as it is linked with variables

suh as temperature, evapora-

tion, louds and aerosols, new

knowledge on the subjet may

improve our ability to under-

stand the limate variations we

experiene, and to predit fur-

therhanges.

ThishapterwillgiveashortreviewofEarth'senergybalaneandthenaturalgreenhouse

eet. I will then try go give an overview of the most ommon auses of natural limate

variations, and briey summarize the ways in whih mankind aets limate. A setion will

be dediated to more reent limate hanges, and last but not least the phenomenon global

dimmingand its onnetionto the observed hangeswill be addressed.

3.1 Earth's energy balane

The main driving fore in Earth's limate balane is the Sun. Every seond, the top of the

atmospherereeives342

W m ⁻²

[2001, Setion 1.2.1℄). This is illustrated in Figure 3.2 on the next page. Of the remaining

235

W m ⁻²

W m ⁻²

Figure 3.2: Global mean energy ows between the surfae and atmosphere [Harvey, 2000,

Figure 2.1℄.

warms the Earth'ssurfae, whih returnsthe heatbak to the atmosphere partlyasinfrared

radiation, partly as sensible heat and partly as water vapor whih releases its heat when it

ondenseshigher upinthe atmosphere(latent heat). Inorderto maintainan energybalane,

the limate system must radiate bak into spae the same amount of energy that entered it

(that is, 235

W m ⁻²

Stefan-Boltzmann's law 1

,the Earthshouldin orderto radiate235

W m ⁻²

◦

C. Yet under present onditions the global mean surfae temperature is 14

◦

C .

The33

◦

Cdierene hasto do with the natural greenhouseeet, brieyexplainedbelow.

3.1.1 The natural greenhouse eet

Justliketheatmosphere absorbssomeofthe inomingsolar radiation,italsoabsorbsinfrared

radiation emitted by the Earth's surfae, gases, partiles and louds. The absorbed infrared

radiationis re-emitted in all diretions, inluding bak downto the surfae. The part ofthe

radiation that is re-emitted out to spae is now being emitted from a muh higher altitude

with temperatures of, on average, -19

◦

C, whih brings the amount of outgoing radiation in

balane with the inoming. This whole proess, where heat is trapped within the system

instead of esaping out to spae, is what we all the natural greenhouse eet. Atmospheri

gasesthatabsorbinfraredradiationisthusalled greenhousegases (oftenabbreviated GHGs)

as they ontribute to the warming, the most eient partiipants being water vapor, arbon

dioxide andmethane.

3.2 Natural limate hanges

In spite of the balane desribed above, the state of the limate is onstantly undergoing

hanges. Suh limate variations our over time sales from a few years to hundreds of

millionsofyears,andresultfromaombinationofperiodiforingmehanisms(bothexternal

and internal), and a omplex group offeedbakmehanisms thatoperate within the limate

systemitself. Hereafterfollowsanaount ofthemainexternal andinternalauses oflimate

hange.

1

Stefan-Boltzmann's law is given by

F

ǫσT ⁴

F

ǫ

T

σ

3.2.1 External auses of limate hange

External foring mehanisms involve agentsor proesses atingfrom outsidethe limate sys-

tem. Theyaet the limate system, but arenot aetedby it.

The Sun with its yles of intensity is an external foring agent. Currently we know of

a2,400yearyle,a200yearyle,an80to90yearyleandtheshorter11and22year

yles [Buhdahl, 1999℄. Obviously, the more ative the Sun is, the more radiation it

emitsand themore isreeived bytheEarthsystem. Attempts have been madetodraw

onnetions between the number ofsunspotson the Sun's surfae and the temperature

on Earth: As the sunspots are areas of ooler temperature in the Sun's photosphere,

a high number of sunspots will redue the average energy emission from the Sun and

onsequently the amount of energy that reahes the Earth. However, it is diult to

diretly attribute an observed limate hange to variations in solar irradiane, beause

thelatteraresmallinmagnitude. Thesolarirradianegenerallyvariesbylessthan0.1%

over the ourse of asunspot yle[Hartmann,1994, p.289℄. Withsuhsmall variations

in thesolar onstant,the globallimatiresponsewouldbenomore thana0.03

◦

Ctem-

perature hange[Buhdahl, 1999,Setion 2.5.3℄. Moreover, ahange in solarirradiane

willindueahangeintheamountofozoneproduedintheatmosphere,whihworksas

anegativefeedbakanddampenstheinitial radiationhange [Harvey,2000,p.29℄. Still,

Friis-Christensenand Lassen[1991℄, whonda goodorrelationbetween the variations

in the solar ylelength and the variations in Northern Hemisphere landtemperature,

onludethatthesolarativitydoeshaveadiretinueneongloballimate. Aphysial

mehanismouplingthe twowashowevernot given,but thenext paragraphmaygivea

suggestion of one:

Cosmi ray variation hasafter the 1997 artile of Svensmark and Christensen been spe-

ulated to have an inuene on limate through aeting the amount of low louds and

hene Earth's reetivity. The intensity of the osmi ray ux is inversely related to

the sunspot yle, so that in periods of high solar ativity (and hene few sunspots),

the Sun worksas a shieldproteting the Earthfrom osmi rays from the surrounding

galaxies[Carslawet al.,2002℄. Oppositely,whentheSunislessative,moreosmirays

ndtheirwaytoouratmosphere. Theosmiray-loudtheorylinkstheionizingosmi

raystothe formationofsulphateaerosols,whihanfuntionasloudondensationnu-

lei andtherebyaet theformation oflouds[Svensmark andFriis-Christensen,1997℄.

This means that periods of lowsolar ativity (high osmi ray ux) will be assoiated

with inreasedloudover,inreasedEarthreetivityandultimatelyloweringthetem-

perature. Svensmark and Friis-Christensen [1997℄ do in fat relate the observed 3-4%

variationofgloballoudoverduringthereentsolaryletovariationsintheosmiray

ux. Thetheoryisbeingopposedbymanydueto lakofaproper physial mehanism

to linkthe ionized partileswith the formation oflouds nearthe surfae [Kristjànsson

and J.Kristiansen, 2000; Sun and Bradley, 2002℄, and due to the fatthat the original

graphsofFriis-ChristensenandLassen[1991℄ aordingto[DamonandLaut,2004℄have

been subjetto inorret data proessing. Although an attempt was madebyCarslaw

et al.[2002℄ to improvethe physial basisof the theory, the questionof whetheror not

variations in the osmi rayuxinuene limateremains unanswered.

The Milankovith theory providesanotherontributiontonaturallimatehange. Earth's

limate isvery dependent on itsposition in and properties of its orbit around the Sun.

Variations inthe eentriityofthe Earth'sorbit(dominant period: 100,000years),the

obliityofthe Earth'saxis(period: 41,000years) andthe preessionof theEarth'saxis

(period: 23,000years)arethe threeyles inludedin the Milankoviththeory[Skinner

Eah ylehasits ownwayof regulatingthe amount anddistribution ofsolar radiation

reahing our planet, ausing long-term variations of as muh as 10% in the amount of

radiant energy reahing Earth [Skinner and Porter, 2000, p.335℄. The eets of eah

yle's phase an either enhane or weaken eah other, and bysumming the variables

we an derive the ombined signal (with variations up to 25%) desribing the amount

of radiation reeivedat anygiven latitude at anyspeied time in the past. Aording

to the Milankovith theory itis the preession of these yles thatontrols the timing

of the 100,000 yearieage yles.

The movement of ourSolar system through the galaxy isalso speulated to have an

eet of limate [Buhdahl, 1999, Setion 2.5.3℄. Passing through denser and lighter

louds ofinterstellar mediawill regulatethe amount ofsunlight thatreahesthe Earth,

and variations in the gravitational torque indued by nearby Magellani louds ould

also somehowhave onsequenesforlimate. However,dueto theenormoustime sales

assoiated with this type of foring, it is impossible to prove these speulations with

observations.

3.2.2 Internal auses of limate hange

Changesin limate an also be triggered by variations within the limate system, fored by

mehanismsoperating overtimesalesof

1

10 ⁸

long-term internal variations isthe globaldistribution ofontinents,whih isdrivenbyinter-

nal plate tetonis moving lithospheri plates about the globe at a rate of a few entimeters

per year[Skinner and Porter, 2000, p.18℄. Also aeting limate is the formation and distri-

bution ofmountain ranges,whihespeiallyinuenesthehorizontal atmospheriirulation

patterns. Changesintheseomponentsofthe lithospherearehowever soslowthattheyoften

are regarded asonstant.

Changes in oean irulation patterns ontribute to limate hange. Espeially impor-

tant isthe thermohaline irulation, where warm surfae water moving northward into

the NorthAtlantievaporates, inreasingthe salinityofandooling the underlyingwa-

ter, whih then sinks and turns bak south [Skinner and Porter, 2000, p.337℄. As the

waterools,heatisreleasedto the atmosphere,maintaining a relativelymild limatein

north-westernEurope. Changesbetweentheglaialandinterglaialpatternsofthether-

mohalineirulationisanatural partofEarth'sshiftinglimateandan beregardedas

internal limate foring mehanisms; for even if irulation hanges are initially driven

by hanges in limate, oean irulation has traditionall y been viewed as an internal

foring mehanism inits ownright [Buhdahl,1999℄. Thesubjetisalso reeivingalot

of attention in onnetion with reent limate warming, as a weakening of, hange in

or omplete shutdown of the thermohalin e irulation will have substantial impat on

European limate.

Climati impats of major volani eruptions generallylastforshorttimeperiods(typ-

ially 2 to 3 years), but the ejeted volani dust and gases an lower average surfae

air temperature s by as muh as 0.5 to 1.0

◦

C in that period [Skinner and Porter, 2000,

p.337℄. Inaseofextremelylargeeruptions, thelimatieetsouldbemoreextensive,

asisbelievedbysometohavebeentheaseinTheLittleIeAge(1300to1850)[Robok

et al.,2000℄. Roboket al. [2000℄ also point out thatperiods of little volaniativity

an ause a warming as this givesthe atmosphere time to washout reetive volani

aerosols. Thiswasprobablytheaseforthe50yearsfrom1912to1963. Itisadditionally

made from outgassing from volanoes, but this extremely slow hange in atmospheri

ompositiondue to volanoesisnot traditionally thought ofaslimate hange.

Changesin atmospheri omposition for other reasons are however a very important

ause of limate hange. Although we usually hear of hanging gas onentrations in

onnetion with anthropogen i emissions, the atmospheri omposition also hanges

naturally. This has been shown through studies of Antarti ie ores. During glaial

times, theperentagesofarbondioxideandmethane seemtobelowandthe amountof

windblowndustunusuallyhigh[SkinnerandPorter,2000,p.336℄. Changesintheamount

and distribution of vegetation as temperatures hange will also have an impat on the

amount of arbon dioxide in the atmosphere. As the atmospheri omposition varies,

the absorbingand reetingpropertiesof theatmosphere hange orrespondingl y. This

largely aets the radiation budget and hene atmospheri and surfae temperatures.

Similar eets an additionally our from variations in the amount of aerosols and

louds, whihwasdisussedin Setions 2.2and 2.3respetively.

3.3 Anthropogeni limate hanges

Humanbeingsaet limatein manyways.

Fromearlytimeswestarteduttingdowntrees

andgrowingelds,hangingthevegetationand

landsape after our needs. This is happen-

ing to a larger and larger extent today; green

landsapes are replaed by asphalt roads and

onrete ities, vast areas of forest are burnt

downandequallylargeareas areploughed and

turnedintoelds. Thesehangesinsurfaeap-

pearane result in hanged surfae albedo and

alsoaet the properties ofthe soil. Thismay

in turn lead to spreading deserts near defor-

estedareas, hanges in the gases released from

thesoilorvegetation,andinreasedwindblown

dust, whih both loally and regionally an

ausesubstantial hangesin limate.

Figure3.3: Skethfrom1858ofaskeletonsymboliz-

ingpollutionintheThamesasresultoftheindustrial

revolution[Aastad, 1997,p.25℄.

GasesandpartilesarereleasedtotheEarth-atmospheresystemthroughindustry, transport,

agriulture, and biomass burning. A large portion of the emittedgases (namely arbon diox-

ide,methaneandnitrogen)areeientgreenhousegases,some(primarilyCFCgases)damage

the ozone layer, while others (like sulphate dioxide) an onvert into aerosols, aeting the

radiationbudgetdiretly throughreetion andabsorption, or indiretly byalteringthe life-

timeand properties oflouds. Agreatnumber ofaerosols arealsoemittedto the atmosphere

diretly, as was mentioned in Setion 2.3 on page 10. The next setion will deal with the

observedhangesin limateoverthe lastfewdeades. Some, likethehange intheamountof

greenhouse gases and aerosols in the atmosphere, are obviously anthropogen i, while others,

likehangesin loudamountsor in atmospheriirulation, mayormaynot be onneted to

humanativities.

While the debate on global warming goes on, and the role of mankind in the hanging

limate is onstantly disussed, itishardlydeniable that tosome extent we are beginning to

set our markonthis systemin whih we reside.

3.4 Reent hanges in numbers and gures

Surfae temperatures have, aording to IPCC [2001, Setion 2.3℄, inreased by

about 0.6

◦

Csine thelate 19thentury(see

Figure 3.4). Although global, the warm-

ing is not entirely homogeneou s. For ex-

ample, sine the mid-1980s Northern Hemi-

sphere land surfaes have warmed onsid-

erably faster than those on the Southern

Hemisphere [IPCC, 2001, Setion 2.2.2.1℄,

and some regions even experienes oppo-

site trends. An example of this is given by

Tuomenvirtaet al.[2000℄,thatreportofde-

reasingtemperature sin westernGreenland

sinethe1950sonnetedwithastrengthen-

ingof the NorthAtlanti Osillation. IPCC

[2001, Setion 2.2.2.1℄ point out that land

Figure3.4: Temperatureanomaliesfrom1860to2000,

IPCC[2001,Figure2.1℄.

surfae warming was twieas fast between 1975 and 2000 asbetween 1920 and 1945, whih

indiates that the rate of warming is also inreasing. Additionally, the warming seems to

involve a faster rise in daily minimum temperatures than in daily maximum temperatures:

from 1950 to 1993 the maximum temperature s inreased by 0.1

◦

C/deade while minimum

temperatures inreased by0.2

◦

C/deade [IPCC, 2001, Setion 2.2.2.1℄. Thismeans thatthe

diurnal temperaturerange(DTR)isdereasing,diminishingthedierenebetweennight-and

daytimetemperature s. Easterlinget al.[1997℄,whofoundsimilarresultsa fewyearsprevious

tothe2001IPCCreport,attributedthedereasingDTRtovariationsinNorthernHemisphere

irulation patterns.

Sea surfae temperatures aswellasthe temperature in the deeper layersof the oean

Figure 3.5: Sea surfae temperatu re anomalies from

1860 to2000, IPCC[2001,Figure2.5℄.

have inreased worldwide. Levitus et al.

[2000℄reportofa0.06

◦

Cwarmingfrom1948

to 1998 ofthe surfae to 3000m layer ofthe

global oeans, anda 0.31

◦

Cwarming ofthe

surfae to 300m layer. The temperature in

the upper few meters of the oean has, as

anbeseeninFigure3.5,shownaninrease

oflose to 0.7

◦

Cover thelast entury. Asa

resultofthermalexpansion(volumeinrease

ofthe oean asitwarms)and iemelting,a

rise in global average sealevel by 1.0to 2.0

mm/yrhasbeenobservedoverthe20then-

tury,andtherearealsoreportsofhangesin

the depth ofthe mixed-layer in severalregions [IPCC,2001, Chapter11℄.

Land preipitation has experiened an overall global inrease of about 2% sine the

beginning of the 20th entury [IPCC, 2001, Setion 2.5.2.1℄. No pronouned hanges are

however observed over Southern Hemisphere land; and China, south-eastern Asia, eastern

Russiaand many equatorial regions have seen slightly dereasing trends. As limate warms

and the water holding apaity of the atmosphere inreases, there is aording to Karl and

Trendberth [2003℄ a tendeny for preipitatio n events to beome more intense. This is for

example the ase in China, in spite of a weakening of the total preipitatio n amount. Even

though therearerelatively fewstudiesof snowfalltrends worldwide, observationsindiate an

inreasein line with the general preipitatio n trend[IPCC, 2001, Setion 2.2.5.1℄.