Climate change and the greenhouse effect

A briefing from the Hadley Centre

December 2005

Cl imate change 1

The Hadley Centre for Climate Prediction andResearch has been a vital department of the MetOffice since it was established in Bracknell in 1990,and continues to be so after its move (with the restof the Met Office headquarters) to Exeter in 2003.Most of its 120 staff give presentations and lectureson all aspects of the research. Many of these are toconferences and workshops on the detailedspecialisms of climate change, such as themodelling of clouds or the interpretation of satellitemeasurements of atmospheric temperatures. Butothers are to less specialised audiences, includingthose from Government — ministers and seniorpolicymakers — industry and commerce, pressuregroups and the media, in each case from the UKand overseas.

In 1999 we collected some of the slides that hadbeen drawn up to give these talks — mostly basedon results from the Hadley Centre — into a generalpresentation, and made this freely available,recognising that this would reach a wider audience.Although our basic understanding of the science ofthe greenhouse effect was well founded in 1999,the vast amount of new research carried out overthe past six years has produced many interestingand exciting results, and strengthened some earlier,more tentative conclusions (such as the attributionof recent climate change to human activities) whichwe decided to reflect in this second edition of thepresentation; in fact, only seven out of the 61 slidesremain from the first edition. This presentation canbe used simply as a self briefing or, better still, as atool to brief others, perhaps by selecting a subset ofthe slides and notes. We encourage their furtherdissemination and use, asking only that the MetOffice Hadley Centre (or another originating institutewhere shown) is fully acknowledged as the source.

I would like to thank Geoff Jenkins for puttingtogether this presentation and the text on the notespages, and Chris Durman, Chris Folland, JonathanGregory, Jim Haywood, Lisa Horrocks (Defra), TimJohns, Cathy Johnson (Defra), Andy Jones, ChrisJones, Gareth Jones, Richard Jones, Jason Lowe,James Murphy, Peter Stott, Simon Tett, Ian Totterdell,Peter Thorne and David Warrilow (Defra), for theircomments, which have greatly improved it. FionaSmith and the Met Office Design Studio handledthe layout and production. Consistent funding fromthe Global Atmosphere Division of the UKDepartment of the Environment, Food and RuralAffairs, and the Government MeteorologicalResearch programme over the last 15 years, hasenabled us to plan a sensible long-term researchstrategy. Finally, I would thank all members of theHadley Centre, and those from other institutes, forletting us use their results here.

Professor John Mitchell, FRS, OBE

Chief Scientist

Met Office

Foreword

2 Cl imate change

The Hadley Centre for Climate Prediction andResearch, a division of the Met Office based atExeter, is the UK Government’s research centre intoclimate change. It was opened by the then PrimeMinister, Margaret Thatcher, in 1990, althoughwork on climate change had been taking place atthe Met Office for a decade or more before that.The aims of the Centre, relevant to climate change,are to:

n understand processes within the climate systemand develop models which represent them;

n monitor climate variability and change on globaland national scales;

n simulate climate change over the past 100 yearsand predict change over the next 100 years andbeyond;

n attribute recent changes in climate to specificnatural and man-made factors;

n predict the impacts of climate change, such asthose on water availability and the capacity forfood production.

The programmes of research are funded mainly bycontracts from the Department of the Environment,Food and Rural Affairs (Defra); the GovernmentMeteorological Research and the European UnionDGXII Climate and Environment Programme,together with smaller contracts from the Foreignand Commonwealth Office.

Many of the slides in this booklet give references inabbreviated form to one of the following reports:

TAR IPCC 2001: Climate Change 2001: TheScientific Basis. Contribution of WorkingGroup I to the Third Assessment Report ofthe Intergovernmental Panel on ClimateChange. Houghton, J.T., Y. Ding, D.J.Griggs, M. Noguer, P.J. van der Linden, X.Dai, K. Maskell and C.A. Johnson (eds).Cambridge University Press, Cambridge,UK, 2001.

4AR Fourth Assessment Report from the IPCC,due to appear in 2007.

SRES Special Report on Emissions Scenarios. ASpecial Report of Working Group III of theIPCC. Nakicenovic et al (eds). CambridgeUniversity Press, Cambridge, UK, 2000.

UKCIP02 Hulme, M., G.J. Jenkins et al. ClimateChange Scenarios for the UnitedKingdom. April 2002.

see: www.ukcip.org.uk/resources/publications/pub_dets.asp?ID=14

1

Climate change

Cl imate change 3

Before looking at how the climate has changedrecently, and the effects of man’s activities, it isworth looking at the natural climate of the Earthand how this can be explained. The map above*shows the long-term average distribution of annualtemperature at the land surface of the Earth,collected from weather and climate observingstations. Because there are so few measurements inAntarctica, it is not possible to colour the map inwith confidence over this continent, although theAntarctic is in fact colder than the Arctic. As werecall from school geography lessons, temperaturesnear the equator are generally high, whereastemperatures near the poles are low, with a gradientin between the two. The reason for this is given inthe next slide. Other factors also affect surfacetemperatures, notably altitude, which explains whymountainous regions, such as the Himalayas, arecolder than other land at the same latitude.

*Taken from the IPCC Data Distribution Centre at the Climatic Research Unit, University of East Anglia:http://ipcc-ddc.cru.uea.ac.uk/java/visualisation.html

2

4 Cl imate change

Because the Earth is a sphere, solar radiation(sunlight) striking it near the equator will be spreadout over a relatively small area, and will have a largeheating effect per square metre. At higher latitudes,sunlight will strike the Earth’s surface moreobliquely, be spread out over a larger area, andhence have a smaller heating effect per squaremetre. This is why equatorial regions are hot andpolar regions are cold. This temperature differencedrives weather, which seeks to minimise thistemperature gradient through the generalcirculation of the atmosphere. One majorcirculation system is that in which air rises near theequator, moves polewards in the higheratmosphere, sinking in higher latitudes, andblowing equatorwards near the surface as the tradewinds. This system is known as the Hadley Cell afterits discoverer George Hadley (1685–1768) — afterwhom the Hadley Centre is also named.

The diagram shows the situation at the equinoxes(21 March and 21 September) when the Sun isdirectly over the equator at noon. During thenorthern summer, the northern hemisphere is tiltedtowards the Sun, making the Sun higher overhead(hence solar heating more intense) and the daylonger. During the northern winter, the northernhemisphere is tilted away from the Sun, so the Sunappears lower in the sky and the day is shorter. Thisgives the seasonal cycle of temperature through theyear.

Climate is the description of the long-term averagesof weather, usually taken over a 30-year period. Itdescribes not only the long period averages oftemperature, rainfall and other climate quantities indifferent months or seasons, but also the variabilityfrom one year to the next.

3

Cl imate change 5

Previous slides discussed the distribution oftemperatures across the Earth, but the Earth alsohas a global average temperature, averaged acrossits entire surface and over a long period such as adecade, of about 14 °C. Two streams of energydetermine what the average surface temperature ofthe Earth is.

Firstly, energy coming in from the Sun as sunlight,which we can see with our eyes. This acts to warmthe surface of the Earth and the atmosphere.

Secondly, because the surface of the Earth (even onthe coldest night in Antarctica) is warmer thanouter Space, infrared radiation is everywhere beingemitted from the surface of the Earth. This acts tocool the Earth’s surface.

We cannot see infrared with our eyes, but it isstraightforward to measure with instruments. In the image of the UK shown above, a satelliteinfrared imager shows cold areas (such as the opensea and cloud tops) as green, blue and purple;warm areas (such as coasts and inland countryside)as yellow and red; and very warm areas (mainlyurban) in black.

The balance between these two streams of energy— that emitted by the Sun and that emitted by theEarth — determines the global average surfacetemperature of the Earth.

If the amount of solar radiation reaching the Earth,or the amount of infrared radiation leaving theground, changes, then the Earth’s temperature willchange. Any change will be very slow because theEarth's climate system has a large thermal inertia,mainly due to the ocean.

Photo credits, top: Space Environment Center, NOAA, USA.

4

6 Cl imate change

In order to understand the greenhouse effect onEarth, a good place to start is in a greenhouse. Agreenhouse is kept warm because energy coming infrom the Sun (in the form of visible sunlight) is ableto pass easily through the glass of the greenhouseand heat the soil and plants inside. But energywhich is emitted from the soil and plants is in theform of invisible infrared (IR) radiation; this is notable to pass as easily through the glass of thegreenhouse. Some of the infrared heat energy istrapped inside; the main reason why a greenhouseis warmer than the garden outside. However, this isa rather crude analogy to the way the greenhouseeffect works on Earth, as will be seen next.

The existence of the greenhouse effect has beenknown about for a long time. In a paper publishedin 1896 by the Swedish scientist Arrhenius, hediscusses the mean temperature of the groundbeing influenced by the presence of heat-absorbinggases in the atmosphere. Indeed, the notion of heatabsorption by gases was put forward even earlierthan Arrhenius' time, by scientists such as Tyndalland Fourier. So the greenhouse effect may berelatively new to policy makers and the media, butin the scientific community it has been knownabout, and investigated, for well over 100 years.

5

Cl imate change 7

As explained earlier, the temperature of the Earth isdetermined by the balance between energy comingin from the Sun in the form of visible radiation,sunlight, and energy constantly being emitted fromthe surface of the Earth to outer Space in the formof invisible infrared radiation. The energy coming infrom the Sun can pass through the clearatmosphere pretty much unchanged and thereforeheat the surface of the Earth. But the infraredradiation emanating from the surface of the Earth ispartly absorbed by some gases in the atmosphere,and some of it is re-emitted downwards. The effectof this is to warm the surface of the Earth and lowerpart of the atmosphere. The gases that do this workin the natural atmosphere are primarily watervapour (responsible for about two-thirds of theeffect) and carbon dioxide.

A rather more rigorous explanation is given in thenext slide.

The natural greenhouse effect has operated forbillions of years. Without the greenhouse effect dueto natural water vapour, carbon dioxide, methaneand other greenhouse gases in the atmosphere, thetemperature of the Earth would be about 30 °Ccooler than it is, and it would not be habitable. Sothe greenhouse effect due to naturally occurringgreenhouse gases is good for us. The concern isthat emissions from human activities (for example,CO2 from fossil-fuel burning) cause thesegreenhouse gas concentrations to rise well abovetheir natural levels and as new greenhouse gases(such as CFCs and the CFC replacements) areadded, the further warming which will take placecould threaten sustainability. This is discussed inlater slides.

Further reading: Houghton, J.T., Global Warming. Rep. Prog. Phys. 68, 1343–1403 (2005).

6

8 Cl imate change

A more rigorous explanation of the greenhouseeffect is as follows. Temperature in the loweratmosphere (troposphere) decreases with height,on average. That for the present day climate isshown by the black line in the above diagram.(Temperature is shown in units of Kelvin (K); 0 °Cis about 273 K). The infrared radiation that coolsthe Earth comes from an average height of about5.5 km at present, due to absorption and re-emission of greenhouse gases. In a future worldwith higher greenhouse gas concentrations,emission of IR to space will be from a higher (andtherefore cooler) layer (in the hypothetical exampleshown above, just over 7 km). Because the rate ofemission of IR increases with temperature, emissionfrom this cooler layer will be reduced, and theatmosphere then warms up (red line) until the rateof IR emission to space reaches the original rate.The surface temperature will then be warmer thanthat in the present day (by about 10 K, that is 10 °C,in the hypothetical figure above).

Further reading: Houghton, JT, Global Warming. Rep. Prog. Phys. 68, 1343-1403 (2005).

7

Cl imate change 9

Before moving on to discuss recent changes andtheir causes, it is worthwhile setting the long-termcontext.

It is well known that global temperatures changesubstantially over timescales of a hundred thousandyears, as climate moves from ice ages to interglacials.The figure above shows measurements* deducedfrom ice cores drilled from the Greenland ice sheetand analysed by the British Antarctic Survey andothers as part of the European programme EPICA(Figure courtesy Eric Wolff, BAS). The actualmeasurement is of the concentration of deuteriumin air bubbles, and this can be related to localtemperatures. The figure shows that temperaturerise between the depth of the last ice age 20,000years ago and the current interglacial (Holocene) isabout 9 °C. Swings between glacial and interglacialclimate are likely to be initiated by subtle differencesin the Earth's orbit and tilt of axis around the Sun,known as the Milankovic Effect after its proposer.

Although these orbital changes dictate that theEarth will enter another ice age, this is unlikely tobe for many thousands of years — quite a differenttimescale to that of man-made warming. Someresearchers believe that man-made warming mayactually prevent the Earth from entering anotherice age.

*Reference: Eight glacial cycles from an Antarctic ice core. EPICA Community Members, Nature, 429, 623–628, 2004.

8

10 Cl imate change

In addition to long-term changes due to the Earth'sorbit, there are two main natural agents which canchange global climate: changes in energy wereceive from the Sun and the effect of volcanoes.

The estimated change in solar irradiance — theamount of energy received at the Earth from theSun — is shown here. There are several estimates ofthis quantity; the one shown here is due to Lean,Beer and Bradley*, updated to 2003. The eleven-year solar cycle is clearly seen, and so too is a risebetween about 1900 and 1960, with little if anychange after that. Solar irradiance before 1978 isestimated from proxy data (sunspots, etc) and isless reliable than that measured since then bysatellites. Based on the Hadley Centre HadCM3climate model, we can estimate the globaltemperature increase which the changing solarradiation may have caused; this is shown on theright-hand scale and amounts to one or two tenthsof a degree, so may explain at least some of theglobal temperature rise observed in the early partof the 20th century.

(Note that, because of the large inertia of theclimate system, global temperature does notrespond significantly to the 11-year solar cycle).

There are some theories that the solar influence onglobal climate could be amplified by an indirectroute, for example involving stratospheric ozone orcosmic rays or clouds. A review** of currentunderstanding was prepared for the Hadley Centreby the University of Reading and Imperial College,London. This concluded that there is someempirical evidence for relationships between solarchanges and climate, and several mechanisms, suchas cosmic rays influencing cloudiness, have beenproposed, which could explain such correlations.These mechanisms are not sufficiently wellunderstood and developed to be included inclimate models at present.

However, current climate models do includechanges in solar output, and attribution analysesthat seek to understand the causes of past climatechange by comparing model simulations withobserved changes, do not find evidence for a largesolar influence. Instead, these analyses show thatrecent global warming has been dominated bygreenhouse gas-induced warming, even when suchanalyses take account of a possible underestimate ofthe climatic response to solar changes by models.

*Lean J., J. Beer and R. Bradley, Reconstruction of solar irradiance since 1610: Implications for climate change,Geophys. Res. Letts. 22, 3195–3198, 1995.** The Influence of Solar Changes on the Earth's Climate. L.J. Gray, J.D. Haigh, R.G. Harrison. Hadley CentreTechnical Note 62, January 2005. This can be downloaded fromwww.metoffice.gov.uk/research/hadleycentre/pubs/HCTN/index.html

9

Cl imate change 11

Volcanoes inject gas into the atmosphere. If they areenergetic enough, this gas will reach the stratosphereand form small sulphate aerosol particles which canpersist for a few years. They reflect back some of thesolar radiation which otherwise would have heatedthe surface of the Earth, and hence act to cool theplanet. The amount of volcanic aerosol in theatmosphere is very variable, indicated by this timeseries of its estimated optical depth*, and thecooling effect that this would have. Althoughenergetic volcanoes were relatively common in thelate 19th century (for example, Krakatoa in 1883)and early 20th century, and there have beensubstantial numbers of energetic volcanoes sincethe 1960s (most recently, Pinatubo in 1991), therewas a period in the 1940s and 1950s when theatmosphere was relatively clear of volcanic aerosol.The amount of climate cooling due to volcanicaerosols would have been quite small in that period.This unusually low amount of volcanic cooling(together with the increase in solar radiation shownin the last slide) may have contributed totemperatures in the 1940s being relatively highcompared to earlier decades. As with solar energy,optical depth due to volcanic aerosols has beenestimated indirectly before about 1983, and henceis less certain.

*Source: Sato, M., J.E. Hansen, A. Lacis, L. Thomason: Stratospheric aerosol optical thickness. At: www.giss.nasa.gov/sata/strataer/

10

12 Cl imate change

In an earlier slide, mention was made of emissionsfrom human activities enhancing the naturalgreenhouse effect. The main gases involved in thisare shown in the table below.

Emissions of carbon dioxide into the atmospherefrom human activities have increased since theIndustrial Revolution, particularly since about 1950.The graph above shows changes in emissions fromsolid fuel (mainly coal), liquid (oil) and gas, and thetotal emissions from burning fossil fuels. Units aregigatons (billion tons) of carbon per year; tocalculate the amount of CO2, multiply by this by 3.7.Smaller sources such as cement production and gasflaring are omitted for clarity, but included in thetotal. (Source: Carbon Dioxide Information andAnalysis Centre, Oak Ridge National Laboratory, USA.See: http://cdiac.esd.ornl.gov/trends/emis/glo.htm)

In addition to the fossil-fuels source, carbon dioxideis also emitted when forests are cleared and burnt.Figures for this source are less accurate, but thebest estimate from IPCC is 1.7 GtC per year duringthe 1980s.

Gas Main sources Effective lifetime

carbon dioxide CO2 fossil fuel combustion, approximatelyland-use changes 100 years

methane CH4 natural gas extraction, 12 yearsagriculture

nitrous oxide N2O fossil fuel combustion 114 years

ozone in lower reactions between 3 monthsatmosphere emissions from transport

and industry

11

Cl imate change 13

Emissions of CO2 from human activities becomeinvolved in the natural carbon cycle, a system offluxes of CO2 between land (vegetation and soils),ocean (water and ecosystems) and the atmosphere.The IPCC TAR estimated that, averaged over thedecade of the 1980s, fossil-fuel burning emitted 5.4 GtC/yr into the atmosphere and land usechange (mainly deforestation) a further 1.7 GtC/yr,giving a total of about 7 GtC/yr. The atmosphereretained about 3.3 GtC of this per year, leading tothe measured rise in CO2 concentration, with about1.9 GtC/yr going into each of the two main sinks,the ocean and the land (soils plus vegetation).Many of these estimates are very uncertain; a moredetailed analysis with uncertainty estimates is givenin Chapter 3, Sections 3.1 and 3.2 of the IPCC TAR.

As can be seen from the slide, these man-madefluxes are much smaller than the fluxes in thenatural carbon cycle. However, the natural carboncycle is in balance, and has led to concentrationsof CO2 in the atmosphere remaining relativelyconstant for the thousand years before theindustrial revolution.

Similar cycles (known by the general term ofbiogeochemical cycles) exist for other greenhousegases such as methane and nitrous oxide. In thecase of methane, emissions from natural andhuman activities undergo complex chemicalinteractions in the atmosphere with species such asthe hydroxyl radical (OH), concentrations of whichare themselves affected by other man-madeemissions, such as carbon monoxide.

Ozone in the lower atmosphere (troposphere) isalso a greenhouse gas. It is formed by atmosphericchemical reactions between man-made emissionssuch as hydrocarbons, nitrogen oxides, carbonmonoxide and methane.

In the case of water vapour, which is the mostimportant natural greenhouse gas, emissions fromhuman activity have very little effect on itsconcentration in the atmosphere, which is mainlydetermined by temperature (and, hence, humanactivity does affect it indirectly, via global warming).

12

14 Cl imate change

The main factors which have caused the rise in CO2

emissions shown in the previous slide are twofold:(a) growth in population (shown in the left-handpanel) and (b) growth in energy use per person(shown in the right-hand panel), as more peopleenjoy a more energy-intensive standard of living,with increased ownership of goods, more servicesand greater travel.

Of course, energy use per person is very differentfrom country to country.

The source of the data used to draw these diagrams was the Open University website:http://www.open.ac.uk/T206/3longtour.htm

13

Cl imate change 15

The concentration of carbon dioxide in theatmosphere was roughly constant at about 280parts per million (ppm) for 800 years (and probablylonger) before the start of the industrial revolution.We know this from analysis of air trapped in bubblesin ancient ice cores in Antarctica and Greenland.These ice-core samples then show a gradual risefrom about 1800, accelerating with time. Directmeasurements of CO2 in the atmosphere have beenmade since the late 1950s, notably by Keeling onMauna Loa in Hawaii. These measurements haveshown a steady rise up to a current annual-averagevalue (2004) of 378 ppm. The figure above comesfrom the IPCC TAR, Figure 3.2.

In the inset we show CO2 concentrations over thelast 10 years at the Mace Head station in Galway,Ireland*. In addition to the trend, the effects ofseasonal cycle of vegetation growth and decay canbe clearly seen.

Concentrations in the atmosphere of othergreenhouse gases have also risen due to humanactivities. Methane was about 800 parts per billion(ppb) 200 years ago, and is now at more than 1750parts per billion, although its rise has levelled off,possibly due to reduction in natural gas leakages.Nitrous oxide has risen from a pre-industrialconcentration of about 270 ppb to a current levelof over 310 ppb. Ozone in the lower atmospherehas a less robust observational long-term record butappears to have increased over the same period byabout 30%.

Reference: A burning question. Can recent growth rate anomalies in the greenhouse gases by attributed to large-scale biomass burning events? P.G. Simmonds, A.J. Manning, R.G. Derwent, P. Ciais, M. Ramonet, V. Kazan, andD. Ryall (2005) Atmospheric Environment, 39, Issue 14, pp. 2513–2517.Further reading: IPCC TAR Chapter 3. *Laboratoire des Sciences du Climat et de l’Environnement, France; National University of Ireland, Galway; UKDepartment for Environment, Food and Rural Affairs.

14

16 Cl imate change

How do we know that the rise in carbon dioxideconcentrations since the Industrial Revolution is dueto emissions from man's activities? The carbon inCO2 has several different forms; the most common(about 99%) is called 12C, but there is a very smallfraction of 14C, which is radioactive, with a half lifeof about 5,700 years. Because fossil fuels are so old,the 14C in them has decayed, so the CO2 given offwhen we burn them has very much less 14C in it. Sothe amount of 14C in the air is being diluted by CO2

emissions from burning coal, oil and gas, known asthe ‘Suess effect’. We can estimate the change in14C in the air from 1850 to 1950 by measuring it intree rings; this estimate is shown above in green.When we calculate what this should be, based onman-made CO2 emissions, the calculation (red line)agrees well with the measurements. This is proofthat the rise in CO2 concentration is due to fossil-fuel burning. The technique fails to work afterabout 1950, because radioactivity from atomicbombs corrupts the technique.

There is other supporting evidence, such as theconsistency between the rise in concentration(unprecedented over the last several hundredthousand years) and man-made emissions, and thenorth-south gradient of CO2 concentration.

Source (tree ring observations): Damon, P.E., Long, A., and Wallick, E.I. (1973) On the magnitude of the 11-yearradiocarbon cycle. Earth Planet. Sci. Lett. 20, 300–306. Source (theoretical calculations): Baxter, M.S. and Walton, A. (1970) A theoretical approach to the Suess effect.Proc. Roy. Soc. London A318, 213–230.See SCOPE 13: The global carbon cycle. Chapter 2: Variations in atmospheric CO2 content, by H.-D.Freyer. http://www.icsu-scope.org/downloadpubs/scope13/chapter03.html#abs

15

Cha

nge

in 1

4 C (

%)

Cl imate change 17

The long, effective lifetime of carbon dioxide meansthat its concentration in the atmosphere would bevery slow to respond to a reduction in emissions.This figure shows the concentration of carbondioxide in parts per million from 1990 to the end ofthe next century. The red line shows that if futurecarbon dioxide emissions follows one ‘business asusual’ projection, then its concentration in theatmosphere will roughly double over the next 100 years.

If we were able to stabilise global emissions atconstant 1990 levels, carbon dioxide concentrationin the atmosphere would still go on risingsubstantially (blue line). And even if emissions werecut in half overnight and continued at that level for100 years, then carbon dioxide concentrationswould still actually creep up (green line). Only by areduction of about 70% in carbon dioxide emissionswould we be able to stabilise its concentration in theatmosphere. That is not the same thing as calling fora cut of 70%; it is simply pointing out that thescience of the carbon cycle leads to that conclusion.

However, the carbon cycle model used in thecalculations above assume no feedback betweenthe climate and the carbon cycle. Slide 60 showsthat the two are actually closely coupled, and theresulting feedback may mean that the reduction inemissions required to stabilise carbon dioxideconcentrations in the atmosphere would be evenlarger than 70%.

16

18 Cl imate change

We have already discussed CO2, methane, nitrousoxide and ozone as greenhouse gases. Othergreenhouses gases such as chlorofluorocarbons(CFC) damage the ozone layer and emissions have,hence, been virtually eliminated as a result of theMontreal Protocol. However, because they havelifetimes of around 100 years, their concentrationsin the atmosphere are only now slowly starting todecrease.

Equal amounts emitted of each greenhouse gas hasa different capacity to cause global warming. Thisdepends upon its lifetime (the longer the emissionremains in the atmosphere, the more time it has toexert a warming influence), the amount of extraoutgoing infrared radiation it will absorb in theatmosphere, and its density. The future warmingeffect, usually taken over the next 100 years, of anextra 1 kg of a greenhouse gas emitted today,relative to 1 kg of CO2, is known as its GlobalWarming Potential (GWP). Current estimates are:methane: 23; nitrous oxide: 296; CFC12: 7300; SF6:22200. CO2 = 1, by definition.

The warming effect over the next 100 years ofcurrent emissions of the greenhouse gases willdepend upon the amount of each gas beingemitted globally and its GWP. When thiscalculation is done, it is seen in this slide thatcarbon dioxide will be responsible for about two-thirds of the expected future warming. About aquarter of the warming is expected to be due tomethane, with other greenhouse gases making upthe rest. Carbon dioxide is certainly the mostimportant man-made greenhouse gas, presentlyand in the future. Note that tropospheric ozone isnot included in this calculation.

Further reading: IPCC TAR Chapter 6, Section 12.

17

Cl imate change 19

Very small particles, known as aerosol, have asubstantial effect on climate. Sulphate aerosol inthe lowest part of the atmosphere (the boundarylayer) is created when sulphur dioxide, emitted byhuman activities such as power generation andtransport, is oxidised. Sulphate aerosol particlesscatter some sunlight, which would otherwise reachthe surface of the Earth and heat it, back out toSpace. They therefore have a cooling influence onclimate. The amount of sulphate aerosol in theatmosphere has increased by three or four timesover the past 100 years or more.

Sulphate aerosol particles also have a further,indirect, effect on climate, shown on the right ofthe diagram. Clouds are generated when airbecomes saturated with water vapour and watercondenses onto small particles (cloud condensationnuclei) to form cloud droplets which reflect somesunlight. In a polluted lower atmosphere, becausethere are more cloud condensation nuclei, then forthe same amount of water we get clouds which havea larger number of smaller droplets. These have agreater surface area and, hence, will reflect backmore sunlight than clouds in clean air. This gives anadditional, indirect, cooling effect of aerosols.

The story is further complicated by the ability ofaerosols to change the lifetime of the cloud. Smallerdroplets are less likely to coalesce to form drizzledrops. So clouds in a polluted atmosphere willpersist longer than their clean-atmosphereequivalents, and the consequent greater cloudcover will also exert a cooling influence.

Other types of aerosol particles can have differenteffects on climate. Black carbon (soot) emitted from(incomplete) burning of fossil fuels, such as indiesel smoke, will absorb solar radiation and causethe atmosphere to heat up. IPCC estimates that thismay have caused about 7% of the man-madeheating effect since pre-industrial times. Windblowndust, inorganic carbon and sea-salt aerosols are alsoimportant.

Aerosols in the boundary layer have a lifetime ofonly a few days, since they can easily be washedout by rain or incorporated into clouds whichsubsequently rain.

Further reading: IPCC TAR Chapter 5.

18

20 Cl imate change

The amount of sulphate (SO4) aerosol in theatmosphere, averaged over the decade of the 1990sand calculated by the latest Hadley Centre climatemodel (HadGEM1), is shown in this slide. Sourcesconsidered are sulphur dioxide from humanactivities, ‘background’ non-explosive volcanoes,and natural di-methyl sulphate (DMS) from oceanplankton. The model takes these, considers theeffect of chemical processes in the atmosphere andphysical processes such as dry and wet deposition,and deduces the resulting concentration of sulphateaerosol, expressed here as a column burden inmilligrams per square metre.

Because aerosols only remain in the atmosphere fora few days, the highest atmospheric concentrationsare estimated to be downwind of the greatestemissions areas in Asia. DMS is an importantcontributor only to oceanic areas away from humanactivities. This uneven geographic distributionmeans that sulphate aerosols have a complex effecton climate, both locally and globally.

Sulphur dioxide gas and sulphate aerosols in theatmosphere can have impacts on human health, andalso lead to acid rain which can fall at considerabledistances from the emissions, acidifying waters suchas lakes and endangering ecosystems. For thesereasons, sulphur emissions have been greatlyreduced in the US and Europe since the 1980s. It isexpected that similar considerations will, in duecourse, lead to reductions of sulphur emissions inAsia and other rapidly developing parts of theworld. When this happens, the current coolingeffect of sulphate aerosols will be reduced,producing a warming effect which would add tothat from greenhouse gases.

19

Cl imate change 21

The external agents which act to change the climateof the Earth, such as greenhouse gases and solarradiation discussed in earlier slides, are known asforcing agents. The change in the energy availableto the global Earth-atmosphere system due tochanges in these forcing agents is termed theradiative forcing of the climate system and has unitsof watts per square metre (Wm-2). Thus, theradiative forcing is an index of the relative globalmean effect of various agents on the climate of theEarth's surface and lower atmosphere.

This slide, taken from the IPCC TAR, shows thechange in radiative forcing over the period 1750 to2000, due to a number of forcing agents, each ofwhich is linked to human activity (except for solarradiation). In some cases (for example, troposphericozone) a best estimate is shown, together with avertical error bar showing the range of estimates. Inother cases, such as mineral dust, only a range ofuncertainty can be given. The level of scientificunderstanding of each of the factors is shown alongthe bottom of the diagram. Man-made changes ingreenhouse gas concentrations represent the biggestand best-understood effect on climate over theperiod, as shown on the far left of the diagram, andcarbon dioxide is the biggest contributor to this.

Further reading: IPCC TAR Chapter 6, Section 13.

20

22 Cl imate change

How quickly the climate will change in the futuredepends upon two factors: how much greenhousegas emissions grow, and how sensitive the climatesystem is to these emissions. We predict futureclimate change in a number of stages, shown in thisfigure. The first thing we need to estimate is thefuture emissions of greenhouse gases and othergases which affect climate change. Theseprojections are deduced from separate modelswhich take into account population growth, energyuse, economics, technological developments, andso forth. We do not carry out this stage at theHadley Centre, but we take future scenarios of theseemissions from others, particularly the IPCC in itsSpecial Report on Emissions Scenarios (SRES).

Having obtained projections of how emissions willchange, we then calculate how much remains inthe atmosphere, i.e. what future concentrations willbe. For CO2, this is done using a model of thecarbon cycle, which simulates the transfer of carbonbetween sources (emissions) and sinks in theatmosphere, ocean and land (vegetation). For gasessuch as methane, we use models which simulatechemical reactions in the atmosphere. Next we haveto calculate the heating effect of the increasedconcentrations of greenhouse gases and aerosolparticles, known as radiative forcing. This is relativelystraightforward because we know their behaviourquite well from laboratory studies. Finally, the effectof the changed heating on climate has to becalculated. This complete pathway, from emissionsto concentrations to heating effect to climate change,can be done within the climate model, describedshortly, which can predict changes in spatial patternsof climate quantities such as temperature at theEarth’s surface and through the depth of theatmosphere and oceans.

The additional heating of the climate system whichwould occur if the concentration of CO2 in theatmosphere was doubled, is about 3.8 Wm-2. In asimple world this would ultimately warm thesurface by about 1 °C. The prediction of climatechange is complicated by the fact that, once climatechange starts, there will be consequences (feedbacks)in the climate system which can act to either enhanceor reduce the warming. For example, as theatmosphere warms it will be able to ‘hold’ morewater vapour. Water vapour itself is a very powerfulgreenhouse gas, so this will act as a positive feedbackand roughly double the amount of warming.Similarly, when sea ice begins to melt, some of thesolar radiation which would otherwise be reflectedfrom the sea ice is absorbed by the ocean, and heatsit further; another positive feedback. On the otherhand, when carbon dioxide concentrations increasein the atmosphere then it acts to speed up thegrowth of plants and trees (the fertilisation effect)which in turn absorb more of the carbon dioxide;this acts as a negative feedback. There are many ofthese feedbacks, both positive and negative, manyof which we do not fully understand. This lack ofunderstanding is the main cause of the uncertaintyin climate predictions; this applies in particular tochanges in clouds which we will return to later.

Following on from the climate change prediction,the impacts of climate change, on socio-economicsectors such as water resources, food supply andflooding, can be calculated. These is usually doneby supplying climate change predictions to other,off-line, impacts models, and Hadley Centre datahave been used by hundreds of impacts researchersin this way. At the Hadley Centre we are alsoincorporating some impacts models into the climatemodel itself, as this has many advantages.

21

Cl imate change 23

In order to estimate climate change, we have tobuild a mathematical model of the completeclimate system. Firstly, the atmosphere; the way itcirculates, the processes that go on in it, such as theformation of clouds and the passage of terrestrialand solar radiation through it. Secondly, the ocean,because there is a constant exchange of heat,momentum and water vapour between the oceanand atmosphere and because in the ocean there arevery large currents which act to transport heat andsalt. In fact, the ocean does about half the work ofthe climate system in transporting heat between theequator and the poles. Thirdly, the land, because itaffects the flow of air over it, and is important in thehydrological (water) cycle. In addition, we modelthe cryosphere; ice on land and sea. All of thesecomponents of the climate system interact toproduce the feedbacks which determine howclimate will change in the future.

22

24 Cl imate change

The climate model is a mathematical description ofthe Earth’s climate system, broken into a number ofgrid boxes and levels in the atmosphere, ocean andland, as shown above. At each of these grid pointsin the atmosphere (for example) equations aresolved which describe the large-scale balances ofmomentum, heat and moisture. Similar equationsare solved for the ocean. The atmospheric part ofthe third Hadley Centre coupled ocean-atmosphereclimate model, HadCM3 — many results from whichare shown in this presentation — has a grid of 2.5°latitude x 3.75° longitude, and has 19 vertical levels.The ocean model has 20 vertical levels and a gridsize of 1.25° latitude x 1.25° longitude. In all, thereare about a million grid points in the model. Ateach of these grid points equations are solved everyhalf hour of model time throughout a modelexperiment which may last 250 or, in some cases,1,000 years.

The Met Office currently uses a NEC SX-6/SX-8supercomputer. The HadCM3 model is run typicallyfor 250 years of simulation (1850–2100) takingabout three months’ clock time on one SX-6 node.Several simulations can be run at the same time.Currently, some 300 terabytes (million millionbytes) of data are stored for future analysis. This isexpected to double each year.

The Hadley Centre has just completed developmentof its new climate model, known as the HadleyGlobal Environment Model (HadGEM1). As can beseen above, this has a higher resolution horizontallyand vertically, on land and over the oceans. It alsoincorporates improvements to the representation ofthe dynamics of the atmosphere, and of manyprocesses in the atmosphere and oceans. Becauseof this, it is about 15 times more expensive to runthan HadCM3. Results from this will be included inthe IPCC Fourth Assessment Report (4AR).

Further reading: Gordon C. et al. The simulation of SST, sea-ice extents and ocean heat transports in a version ofthe Hadley Centre model without flux adjustments. Clim. Dyn., 16, 147–168, 2000.Pope V.D. et al, The impact of new physical parametrisations in the Hadley Centre climate model, HadAM3.Clim. Dyn., 16, 123–146, 2000.Johns, T.C. et al, The new Hadley Centre climate model, HadGEM1: Evaluation of coupled simulations incomparison to previous models. J. Clim. (submitted).

23

Cl imate change 25

This graph shows the change in global averagesurface temperature from 1861 (when sufficientobservations were made to form a meaningfulglobal average) to 2004. The individual annualaverages are shown by red bars, and the blue lineshows a smoothed trend*, with changes shownrelative to temperatures over the last decades of the19th century. The observations combine those ofnear-surface air temperatures with those of sea-surface temperatures. Updates to this graph can beseen at the Met Office Hadley Centre website:

http://www.metoffice.gov.uk/research/hadleycentre/obsdata/globaltemperature.html

This time series has been used in all the reports ofthe Intergovernmental Panel on Climate Change(IPCC); it is a joint effort between the Met Office'sHadley Centre and the University of East Anglia'sClimate Research Unit.

Although there is a considerable year-to-yearvariability in annual-mean global temperature, anupward trend can be clearly seen; firstly over theperiod from about 1920–1940, with little change ora small cooling from 1940–1975, followed by asustained rise over the last three decades since then.

1998 was the warmest year in this time series,followed by 2002, 2003 and 2004. All of the ‘top-ten’ years have been since 1990, and all the ‘top-twenty’ warmest years have been since 1981, withthe exception of one (1944).

* The smoothing is done with a 21-year binomial filter, to suppress variations on timescales less than about a decade. Reference: Jones, P.D. et al, Adjusting for sampling density in grid box land and ocean surface temperature timeseries. J. Geophys. Res. 106, 3371–3380, 2001.

24

26 Cl imate change

The global average temperature measurementsshown in the previous diagram are taken fromthousands of weather stations all across the globeon land, from ships at sea, buoys and, more recently,satellites. They represent a substantial effort on thepart of national meteorological services worldwide.They are corrected to minimise any errors due tochanges in measurement practices, and for artefactswhich may result from changes in the nature ofobserving stations (urbanisation, for example).

Shown in this slide are three independentmeasurements of global mean temperature; sea-surface temperature, the air temperature over theland surface and the air temperature over the sea.For most of the time, they agree with each otherreasonably well, which shows that this temperaturerise is real. As can also be seen, warming since the1970s has been more rapid over land than over theoceans, as would be expected from an increasingman-made greenhouse effect.

The amount of warming observed varies considerablyfrom place to place, because natural variability ofclimate can add to man-made warming in someplaces, or subtract from it in others. Local factors(such as aerosol cooling) may also come into play.

That the Earth's surface has experienced a recentwarming is also supported by the widespreadrecession of mountain glaciers over the last fewdecades, and by measurements made at differentdepths in boreholes, which can be used to estimatethe historical rise in temperature.

References: Jones, P.D. and Moberg, A. Hemispheric and large-scale surface air temperature variations: anextensive revision and update to 2001. J.Clim., 16, 206–223, 2003.Rayner, N.A., D.E. Parker, E.B. Horton, C.K. Folland, L.V. Alexander, D.P. Rowell, E.C. Kent and A. Kaplan, Globalanalyses of sea surface temperature, sea ice, and night marine temperature since the late nineteenth century. J.Geophys. Res. 108 (D14), 4407, 2003.

25

Glo

bal t

empe

ratu

re c

hang

e (°

C)

rela

tive

to 1

961–

90

Cl imate change 27

Despite the careful corrections that are made totemperature observations to account forurbanisation, concerns still remain that part of therise in land temperatures seen over past decades isdue to changes in or around observing sites; forexample, where a town is creeping out towards apreviously rural area where the climate observationis made. We know that the climate of an urbanstation differs greatly from the surroundingcountryside; the so-called urban heat island (UHI)effect. In the night, temperatures do not fall asquickly in cites, as the mass of concrete helps toretain heat, and city surfaces cannot radiate heataway as fast as in the country.

The UHI effect is weakened or destroyed when thereis a strong wind. So if the recently observed warmingis an artefact of urbanisation, then one mightexpect the temperature rise on calm nights to showthis increasing urbanisation and, hence, to be morerapid than that on windy nights. Recent HadleyCentre work* has used daily night-time minimumtemperature measurements at more than 250 landstations over most of the world, during the period1950–2000. Data were taken for the top third mostwindy nights and the bottom third least windynights (that is, the most calm conditions). Trends forthese two subsets are plotted in the slide. Although,as expected, minimum temperatures in windyconditions are somewhat higher, the trend is thesame in windy and calmer conditions. This clearlydemonstrates that warming over the past 50 yearshas not been due to urbanisation.

26

*Parker, D.E., Large scale warming is not urban. Nature, 432, 290, 2004.

28 Cl imate change

What are the causes of changes in global meantemperature observed since the early 1900s, shownin red on this slide? As we have already outlined,natural factors include a chaotic variability ofclimate due largely to interactions betweenatmosphere and ocean; changes in the output ofthe Sun and changes in the optical depth of theatmosphere from volcanic emissions.

The Hadley Centre climate model has been drivenby changes in all these natural factors, and itsimulates changes in global temperature shown bythe green band in the slide above. This clearly doesnot agree with observations, particularly in theperiod since about 1970 when observedtemperatures have risen by about 0.5 °C, butthose simulated from natural factors have hardlychanged at all.

27

Cl imate change 29

If the climate model is now driven by changes inhuman-made factors — changes in greenhouse gasconcentrations and sulphate particles — in additionto natural factors, observations (red) and modelsimulation (green) are in much better agreement.In particular, the warming since about 1970 isclearly simulated. Of course, this agreement may,to some extent, be fortuitous, for example, if theheating effect of man-made greenhouse gases andthe cooling effect of man-made aerosols have beenoverestimated. Nevertheless, the ability to simulaterecent warming only when human activities aretaken into account is a powerful argument for theinfluence of man on climate. Since this initial HadleyCentre experiment*, other modelling centres havebeen able to reproduce the same broad conclusion.

In addition to simulating the global meantemperature, the model also simulates the patternof changes in temperature, across the surface of theEarth and through the depth of the atmosphere.These ‘fingerprints of man-made warming’ havebeen compared to observations, providing evenstronger evidence for the majority of the long-termtrend over the last 50 years having been due tohuman activity.

Further work at the Hadley Centre** andelsewhere*** has recently demonstrated thatwarming over individual continental areas in thelast few decades can only be explained if humanactivities are included.

* Stott, P.A., S.F.B. Tett, G.S. Jones, M.R. Allen, J.F.B. Mitchell and G.J. Jenkins. External control of twentiethcentury temperature variations by natural and anthropogenic forcings. Science, 15, 2133–2137, 2000.** Stott P.A., Attribution of regional scale temperature changes to anthropogenic and natural causes. Geophys.Res. Lett., 30 (14), CLM 2.1 to 2.4. 2003.*** Karoly, D.J., K. Braganza, P.A. Stott, J. Arblaster, G. Meehl, A. Broccoli and K.W. Dixon. Detection of ahuman influence on North American climate. Science, 302, 1200–1203, 2003.

28

30 Cl imate change

In addition to the attribution of changes on a globaland continental scale, we can also examine theinfluence of man on specific weather events.

The summer of 2003 was an unusually warm oneover large parts of the continent of Europe. Themap shows a temperature anomaly of three degreesor more compared to the late 19th century overlarge parts of Europe in August 2003. The warmsummer caused great losses in agriculturalproductivity in southern Europe, and during thehottest days there were some 20,000 excess deathsin urban areas such as Paris. Some of the mainimpacts are discussed in Ciais et al (2005).

Using a combination of observations and modelling,recent Hadley Centre/Oxford University research*estimates with high probability that the risk of suchanomalously high European temperatures hasalready doubled due to the effects of humanactivities such as CO2 from fossil-fuel burning.

* Stott, P.A., D.A. Stone and M.R. Allen, Human contribution to the European heatwave of 2003. Nature, 432,610–614, 2004.Ref. Ciais, P.H. et al, Europe-wide reduction in primary productivity caused by the heat and drought in 2003.Nature 437, 529–533 (2005).

29

Cl imate change 31

In Slide 25 we showed how sea-surface temperatureshave risen, but at a slower rate than temperaturesover land areas, in the last 50 years. When theatmosphere is warmed by the greenhouse effect dueto man’s activities, this warmth is then transferredto the sea surface by interaction between the seaand the air above, and thus sea-surface temperaturesalso rise. But the constant vertical mixing of theocean waters transfers this heat deeper and deeperinto the ocean, and dilutes the warming at thesurface. To understand fully changes to the oceanwe need to look at how the amount of heatcontained in the whole ocean has changed. This wasfirst estimated by the US National Oceanographicand Atmospheric Administration*, and is shown inthe black line above.

The Hadley Centre model has been used to try tosimulate these observations, and the results** arealso shown in the same figure. Estimates of how theheat content of the ocean would have changed ifonly natural factors were responsible results in thegreen line above; this is obviously unable to mimicthe observations. The estimate of heat contenttrends using only man-made greenhouse gases (redline) gives a trend in heat content which is toogreat. However, estimates from man-made gas plusaerosol emissions (yellow), or all factors (man-madeand natural, blue) give a much better agreementwith observations, implying that human activitiesmust be largely to blame. The large decadalvariability shown in the observations cannot be

simulated by the models; it may have something todo with the limited number of observations ofocean heat content.

More recent research, undertaken jointly by theHadley Centre and US scientists, has shown thatmodels reproduce the penetration of heat into thevarious ocean basins over the last 50 yearsremarkably well. Again, this agreement withobservations can only be achieved when modelsimulations include emissions from man's activities.This work shows that the observed warming of theworld's oceans, which has a complex structurewhich varies from ocean to ocean, is largely ofhuman origin.

* Levitus, S., J.I. Antonov, T.P. Boyer and C. Stephens. Warming of the world ocean, Science, 287, 2225–2229(2000).** Gregory, J.M., H.T. Banks, P.A. Stott, J.A. Lowe and M.D. Palmer. Simulated and observed decadal variabilityin ocean heat content. Geophys. Res. Lett. 31, L15312, 2004.

30

32 Cl imate change

Temperatures are routinely measured not only atthe Earth's surface and in the oceans, but in theatmosphere too. The two panels above showchanges in temperatures in the lower stratosphere(roughly 12 km–18 km) and the lower troposphere(from the surface to roughly 7 km), relative to theperiod 1981–1990. Atmospheric temperatures aremeasured using two very different techniques:firstly using thermometers carried aloft on routineweather balloons (radiosondes) and, secondly, since1978, by remote sensing from a Microwave SoundingUnit (MSU) carried on a series of satellites.Measurements from both techniques are subject tocorrections and different methods of analysis Theradiosonde data above (black) is from the HadleyCentre analysis known as HadAT2. Two differentanalyses of the MSU satellite data are shown, one(blue) from the University of Alabama at Huntsville(UAH, shown courtesy of John Christy, available atwww.nsstc.uah.edu/public/msu) and one (red)from Remote Sensing Systems (RSS, shown courtesyCarl Mears, available at www.remss.com).

The stratosphere (top panel) has cooled on averageby about 1.5 °C since the late 1950s. This arises fromseveral causes. Firstly, CO2 does not warm thestratosphere as it does the troposphere; instead itradiates away heat to outer Space and cools it. Thus,increases in CO2 will have led to a greater coolinginfluence. Secondly, ozone in the stratosphere isheated by solar radiation and raises the temperature

there; because stratospheric ozone has decreased(due to man-made CFCs) this will also act to coolthe stratosphere. Lastly, some cooling is suspected tobe due to increased concentrations of water vapourin the lower stratosphere, which radiates away heatin the same way as CO2. The long-term coolingtrend has been punctuated by spikes where aerosolfrom volcanoes Agung, El Chichon and Pinatuboabsorbed solar radiation and thus produced strongstratospheric warming for two or three years.

The lower panel shows the change in global-averagelower tropospheric temperature measured byradiosondes and satellite. Also shown in green is thechange in global surface temperature. It can beseen that, since 1978, the longer term trend and thevariability of two of the three measurements oftropospheric temperature agree well, and alsoagree with the surface observations. The alternative(UAH) analysis of satellite data shows substantiallyless warming than at the surface. Climate modelspredict that we should have seen a relativelygreater warming in the troposphere than at thesurface; this potential discrepancy between modelsand observations is not well understood, althoughuncertainty in observations is the more likelyexplanation. The topic of stratospheric andtropospheric temperature measurements iscurrently undergoing a thorough review* in the US,with the involvement of Hadley Centre staff. Theconclusions are due out in early 2006.

* Karl, T.R., S. Hassol, C.D. Miller and W.L. Murray (eds), Temperature trends in the lower atmosphere: steps forunderstanding and reconciling differences. A report by the Climate Change Science Programme and theSubcommittee on Global Change Research, Washington, DC. (in draft).

31

–1.5–1.0

–0.5

0.0

0.5

1.01.5

–0.8–0.6–0.4–0.2

0.00.20.40.60.8

Cha

nge

in te

mp

. (°C

) C

hang

e in

tem

p. (

°C)

Agung El Chichon Pinatubo

HadAT2UAHRSS

Surface

1960 1970 1980 1990 2000

Cl imate change 33

Global temperature is not the only climate indicatorto have changed substantially over the last fewdecades. Here we show the change in maximumannual 5-day rainfall events, an important driver offlooding. The red line shows the least squares fit,which is statistically significant to 5%.

There have also been changes to the distribution ofatmospheric surface pressure (as measured bybarometers), with reductions in the high latitudes ofboth hemispheres, and increases at middlelatitudes. This has affected weather patterns, andregional rainfall and temperature in bothhemispheres in middle and higher latitudes.

Man-made climate change is expected to intensifythe water cycle, with consequent increased rainfallat higher latitudes. This is observed in the northernhemisphere winter half year, and supportingindications come from the increasing amount ofoutflow from Eurasian rivers into the Arctic Seas.Recent Hadley Centre work* shows that this trendin outflow cannot be simulated with a climatemodel if only natural factors are included, but canbe simulated well when human activities are takeninto account.

* Wu P., R. Wood and P. Stott. Human influence on increasing Arctic river discharges, Geophys. Res. Lett.,L02703, 2005.

Further reading: IPCC TAR Chapter 2, Section 7.

Source: IPCC TAR 32

34 Cl imate change

The IPCC Third Assessment Report commented thatthere is now ample evidence to support a majorretreat of alpine and continental glaciers inresponse to 20th century warming. In a fewmaritime regions, increases in precipitation due toregional atmospheric circulation changes haveovershadowed the effect of increases in temperaturein the past two decades, and glaciers have re-advanced.

Other changes to snow and land ice have beenobserved. IPCC reports that there are very likely tohave been decreases of 10% or so in the extent ofsnow cover since the 1960s, and there is very likelyto have been a reduction of about two weeks inannual duration of lake and river ice in mid-to-highlatitudes of the northern hemisphere over the past100–150 years.

The figure above shows the cumulative change inlength of a number of glaciers, estimated by theUniversity of Utrecht*. For the period from 1900 to1980, 142 of the 144 glaciers for which adequateinformation was available decreased in length. Forthe earlier period 1860–1900, 35 out of 36 glaciersretreated, although the total rate of retreat was notas fast as during the 20th century. Because glacierstake time to react to change in temperature, andbecause there are fewer accessible measurementsafter 1980, glacier retreat due to warming observedin the past 15 years is not yet reflected in the curveabove.

* Oerlemans, H. Extracting a climate signal from 169 glacier records. Science 308 675–677 (2005).

33

Cl imate change 35

As the oceans warm, it might be expected that seaice will melt. In the Arctic, after a period of stabilityfrom the beginning of the 20th century, there hasbeen a decrease in the sea-ice extent by about 1 million km2, as shown above*. The blue line is themonthly difference relative to the average for thatmonth for the whole period (1972–2005); the redline is filtered to show longer period changes.**Theextent of sea ice is defined as the area within whichthe concentration of sea ice is greater than 15%.

The decrease of about 2.5% per decade since 1970is very similar to that simulated by the HadleyCentre climate model. Internal variability of theclimate system, and external natural factors (solarand volcanic) are very unlikely to have caused atrend of this size, suggesting that human activitiesare at least in part responsible. Climate models areable to simulate this decrease well, when theyinclude man-made greenhouse gas emissions as afactor, suggesting that human activities arecontributing significantly to this decrease.

There are indications that the thickness of sea ice inthe Arctic has also decreased, though trends aremore difficult to measure.

In the case of Antarctica, there appears to have beenno significant long-term trend in sea-ice extent.

*using a 128 term (128 month) binomial filter. **This graph is based on methodology described in: Rayner, N.A., D.E. Parker, E.B. Horton, C.K. Folland, L.V. Alexander, D.P. Rowell, E.C. Kent and A. Kaplan, Global analyses of sea surface temperature, sea ice, andnight marine temperature since the late nineteenth century. J. Geophys. Res. 108 (D14), 4407, 2003.

34

36 Cl imate change

The longest direct measurements of sea level comefrom tide gauges and, based on their observations,the IPCC TAR estimated that the rate of global meansea-level rise during the 20th century is in the range10–20 cm, with a central value (not necessarily thebest estimate) of 15 cm. Based on longer periodrecords, the average rate of sea-level rise has beengreater during the 20th century than during the19th century. However, no significant accelerationin this rise has been detected during the 20thcentury.

Over the last ice age cycle, sea level has rangedfrom 5 m higher than today's, at the time of the lastinterglacial about 120 thousand years ago, to 120 mbelow today's, at the depth of the last ice 21,000years ago when glaciers were at their maximumextent.

Models best simulate the rise over the last centurywhen they take into account natural and man-madeforcing. Even then, the simulated rate is significantlyless than that observed, suggesting that predictionsin the future may also be underestimated.

The figure above shows relative sea level for thepast 2–300 years from stations in Northern Europe.The scale bar on the left indicates a relative changeof ±100 mm (10 cm). Satellite measurements overthe past decade indicate a rise of 2.5 mm/year inthe global mean, but with large regional variations.For example, the East Pacific shows a fall in sea levelover that period, with other areas showing a rise oftwice the global mean.

Further reading: IPCC TAR Chapter 11, Section 3.2.

35

Cl imate change 37

There appears to be no clear trend in the globalfrequency of tropical cyclones, also known ashurricanes in the North Atlantic and typhoons inthe Pacific. However, research at MIT* has shownthat the total power dissipated by tropical storms inboth these regions, integrated over their lifetimes,has increased markedly since the mid-1970s. Thistrend is due to both longer storm lifetimes andgreater storm intensities, and is highly correlatedwith change in tropical sea-surface temperatures(SSTs). The figure above* shows smoothed changesin the power dissipated by hurricanes (PowerDissipation Index, PDI) in the western North Pacificand North Atlantic areas, integrated over thelifetime of the storm and over its area, compared tochanges in the annual mean sea-surfacetemperature in the Hadley Centre Sea-Ice and SeaSurface Temperature (HadISST) data set, averagedbetween 30° S and 30° N. The units of PDI arethose of energy, with units multiplied by anarbitrary factor to match the same units as thechange in SST, to facilitate the comparison betweenthe two quantities. The power dissipation appearsto have nearly doubled over the past 30 years.Based on theoretical considerations, only part of theobserved increase in power dissipation is likely tobe directly due to changes in SSTs. Other factorswhich are known to influence development arewindshear and the depth of the warm water layer.

A recent survey** has shown that over the last 35years there had been roughly a doubling of thenumber of hurricanes in the two most intensecategories (known as Saffir-Simpson 4 and 5, havingwind speeds over 56 m/s), with the largest increasesin the North Pacific, Indian and Southwest PacificOceans, and the smallest in the North Atlantic.

In both these papers, relatively short periods ofobservational records are used. The role of naturalcycles of hurricane activity, and the balancebetween natural and man-made influences in therecent record, are still a matter of debate. In orderto attribute recent observed changes to humanactivity, in the context of substantial naturalvariability, further research, ideally involving alonger observational record would be required.

There is mounting theoretical and modellingevidence that tropical cyclones will become moreintense, although not necessarily more numerous,in a warmer world.

* Emanuel, K. Increasing destructiveness of tropical cyclones over the past 30 years. Nature 436, 686–688, 2005. ** Webster, P.J., G.J. Holland, J.A. Curry, H-R. Chang. Changes in tropical cyclone number, duration andintensity in a warming environment. Science, 309, 1844–1846, 2005.

36

38 Cl imate change

Global temperature rise is the average of individualchanges at large numbers of observation stations,so it is not surprising that most of these also exhibita warming trend. In the UK, the longest runningseries of temperature observations is from threestations in central England, where daily instrumentaldata have been recorded continuously since 1772.Yearly averages of Central England Temperature(CET) are shown in the slide in blue, together with afiltered trend in red. The data for 2005 to April areshown in green and the Met Office Hadley Centrewebsite shows an up-to-date graph at:

http://www.metoffice.gov.uk/research/hadleycentre/obsdata/cet.html

The sustained rise of about one degree Celsius inCET since about 1980 is noticeable; due more to anincrease in maximum temperatures (about 1.2 °Cover the same period) than night minimumtemperatures (about 0.7 °C). 1990, 1999, 1949 and2002 were joint warmest years on record.Temperature averaged over such a small scale asCentral England has much more variability thanthat averaged over the entire globe, as can be seenin the slide, and hence it is not surprising that yearsfurther back, such as 1949, were very warm.

Changes in UK rainfall climate have also beenevident. The England and Wales Precipitation (EWP)record, reaching back to 1766, shows an increase inwinter rainfall and a decrease in summer rainfall.The character of rainfall has also changed, with atrend over the past few decades to a greaterproportion of winter rainfall falling in heavy events.Updated graphs of EWP can also be seen at the MetOffice Hadley Centre website:

http://www.metoffice.com/research/hadleycentre/obsdata/climateindicators.html#EWPannual

37

Cl imate change 39

Projections of emissions from human activity ofcarbon dioxide, other greenhouse gases, and othergases which can affect climate, have been made bythe IPCC Special Report on Emissions Scenarios(SRES). There are four main ‘marker’ scenarios,labelled B1, B2, A2 and A1FI, which can bedescribed (as in the UKCIP02 report) as LowEmissions, Medium-Low Emissions, Medium HighEmissions and High Emissions, respectively. Eachscenario is based on a ‘storyline’ of how the worldmight develop, in terms of population growth,economic growth, energy use, etc. A shortsummary of the storylines is given in Appendix 5 ofthe UKCIP02 report.

In this slide, projected emissions of carbon dioxidefrom fossil fuel burning are shown for the fourfuture emissions scenarios, together with estimatedemissions from 1850–2000. Similar data are availablefor methane, nitrous oxide and other greenhousegases and their precursors (for example NOx whichis a precursor of ozone) and sulphur dioxide, fromwhich sulphate aerosols are formed.

IPCC stress that it is not possible to attachprobabilities to each of the scenarios, and neithercan they be considered all equiprobable. Work is inhand, however, for example at the US EnvironmentProtection Agency, to develop probabilistic forecastsof emissions.

38

40 Cl imate change

Experiments with the Hadley Centre climate model(HadCM3) have been used to predict climate changearising from each of the four SRES emissionsscenarios, and these are shown here. Unsurprisingly,the lowest emissions scenario gives the lowestglobal-mean temperature rise by 2100, about 2 °C,and the highest emissions scenario gives the greatestwarming, about 5 °C. Interestingly, warming overthe next three or four decades is similar for each ofthe emissions scenarios, even though (as can beseen from the previous slide) they diverge sharplyafter 2000. This partly reflects the long effectivelifetime of CO2 and the large thermal inertia of theclimate system. We are committed to much of therise of the next few decades almost irrespective ofemissions over that period. (Part of the reason isthat the projections of CO2 and sulphate aerosolhelp to offset one another). The other side of thiscoin is that, if we wish to limit climate change in thesecond half of the century, then emissions wouldneed to be controlled early in the first half.

These results are from the Hadley Centre model;results from other models can be somewhatdifferent, and this is discussed in the section onuncertainties. The IPCC TAR found that thewarming predicted between 2000 and 2100, fromthe four emissions scenarios and from nine globalclimate models, ranged from about 1.5 to 6 °C.

Temperatures over land are expected to increaseabout twice as rapidly as temperatures over theocean. The warming by 2100 over land areas, whereof course we live and work, is predicted by theHadley Centre model to be in the range 3 °C to 8 °C.

Although the Earth's temperature has variedconsiderably over the last 1,000 years for naturalreasons, the rise over the next 100 years due tohuman activities is predicted to be very much largerthan natural variability, even with the lowestprojected man-made emissions.

Reference: T.C. Johns, J.M. Gregory, W.J. Ingram, C.E. Johnson, A. Jones, J.A. Lowe, J.F.B. Mitchell, D.L. Roberts,D.M.H. Sexton, D.S. Stevenson, S.F.B. Tett and M.J. Woodage. Anthropogenic climate change from 1860 to 2100simulated with the HadCM3 model under updated emissions scenarios. Clim. Dyn. 583–612, 2003.

39

Cl imate change 41

The climate model predicts not just global-meantemperatures as seen in previous slides, but also thepattern of climate change across the surface of theEarth, and through the depth of the atmosphereand ocean. The above figure shows the rise intemperature of the air near the land surface, andthe temperature of the surface of the sea – these arecommonly referred to as simply ‘surfacetemperature’. The change in surface temperature inthe northern winter (December–February) averagedover the last 30 years of this century (that is, centredon the 2080s) compared to a recent reference period(1961–1990), assuming emissions follow the SRESHigh Emissions scenario, is illustrated. Changes insurface temperature range from yellow, where thereis very little change, to dark brown, where there isgreater warming. It can be seen that land areas are,in general, redder than the ocean areas. This ismainly because as the warming in the atmospherecauses the sea surface to warm, this warmth will bemixed down by turbulence and spread throughdeeper and deeper layers. This acts to keeptemperature rise at the sea surface relatively smallcompared to that over land. Northern high latitudes,where the disappearance of sea ice acts as a positivefeedback, show the greatest warming. On the otherhand, there are one or two areas of the oceans whichshow a minimum in warming, for instance in thesouthern oceans and in the northern N Atlantic. Inthese areas there is a rapid exchange of surface waterwith very deep water; this acts as a heat sink andallows sea-surface temperature to change only slowly.

Reference: T.C. Johns, J.M. Gregory, W.J. Ingram, C.E. Johnson, A. Jones, J.A. Lowe, J.F.B. Mitchell, D.L. Roberts,D.M.H. Sexton, D.S. Stevenson, S.F.B. Tett and M.J. Woodage. Anthropogenic climate change from 1860 to 2100simulated with the HadCM3 model under updated emissions scenarios. Clim. Dyn. 583-612, 2003.

°C40

42 Cl imate change

Shown in this slide are predicted changes inprecipitation (rainfall and snowfall) under the sameHigh Emissions scenario as the previous slide, andthe same season (northern winter) and time frame(2080s). The red areas are those where rainfall ispredicted to be smaller and the blue areas where itis predicted to be greater; changes in rainfall areseen to be both positive and negative. Mostmodels find that global average precipitationincreases with time, as the hydrological cycle isenhanced by global warming. It changes most inhigh latitudes and in the Indian monsoon, andchanges least in the subtropics. As seen above, theHadley Centre model predicts substantial changesin rainfall in December-February over northernBrazil, for example.

Confidence in specific areas of rainfall change fromclimate models is not as great as the confidence intemperature change. Although all the main modelssee large changes — positive and negative —particularly in the tropical regions, they do notalways agree exactly where those changes will be.

Reference: T.C. Johns, J.M. Gregory, W.J. Ingram, C.E. Johnson, A. Jones, J.A. Lowe, J.F.B. Mitchell, D.L. Roberts,D.M.H. Sexton, D.S. Stevenson, S.F.B. Tett and M.J. Woodage. Anthropogenic climate change from 1860 to 2100simulated with the HadCM3 model under updated emissions scenarios. Clim. Dyn. 583–612, 2003.

mm/day41

Cl imate change 43

Here, we focus on simulation and prediction (undera Medium High Emissions scenario) of summerwarming over southern Europe, using the HadleyCentre model, shown in red. The simulation from1900 to 2000 agrees well with the observations(black), except for the summer of 2003 (blackasterisk) which was much warmer than either themodel simulation or the climatic norm (see Slide29). In the absence of any human modification ofclimate, temperatures such as those seen in Europein 2003 are estimated to be a 1-in-1,000 year event.Despite this, it is seen that, by the 2040s, a 2003-type summer is predicted to be about average, andby the 2060s it would typically be the coldestsummer of the decade.

Recent work has demonstrated that hot, 2003-type,European summers would already be beingexperienced much more frequently (more thanonce per decade on average) were it not for thecooling effect of man-made sulphate aerosols. Oncethe shielding effect of these aerosols is removed, thewarming commitment from past greenhouse gasemissions will mean that such hot summers arelikely to become a regular occurrence, even withoutany further man-made greenhouse warming.

Reference: Stott, P.A., D.A. Stone and M.R. Allen. Human contribution to the European heatwave of 2003.Nature, 432, 610–614, 2004.

Tem

per

atur

e ch

ang

e (°

C)

42

44 Cl imate change

We saw in Slide 34 how the extent of Arctic sea icehas diminished since about the mid-1970s, and howthis can only be replicated by models when man-made greenhouse gas emissions are factored intothe simulations. Using the same Hadley Centremodel, we can predict how Arctic sea ice will changein future. Under the High Emissions scenario wefind that ice in the month of September (when it isat its minimum extent in the annual cycle) will havealmost completely disappeared on average by the2080s. Other emissions scenarios lead to a slowerreduction in Arctic ice. Reductions are alsopredicted for other seasons.

Reference: Gregory, J.M., P.A. Stott, D.J. Cresswell, N.A. Rayner, C. Gordon and D.M.H. Sexton. Recent and futurechanges in Arctic sea ice simulated by the HadCM3 AOGCM. Geophys. Res. Lett, 29, (2002).

43

Cl imate change 45

As seen earlier, global climate models generallyhave a resolution of about 300 km, and this isinsufficient to calculate reliably the impacts ofclimate change. To achieve a higher resolution we‘downscale’ global model predictions using RegionalClimate Models (RCM), which have a resolution of25 or 50 km, but over a sub-global domain, typically5,000 km x 5,000 km. RCMs not only give greatergeographical detail, but they take better account offeatures such as mountains and coastlines, and givea much improved simulation of (and, hence, can beexpected to give better predictions of) changes inextremes such as heavy rainfall events. The accuracyof predictions from RCMs is, of course, limited bythe accuracy of the driving predictions that areinput from global models.

The Hadley Centre RCM (HadRM3) has been usedto predict changes at a resolution of 25 km over aEuropean domain by the 2080s under a Medium-High Emissions scenario, and two results are shownabove. In the left-hand panel we see the predictedpattern of change in summer-mean rainfall,showing almost all of the continent (apart fromnorthern Fennoscandia) becoming much drier. Inthe right-hand panel we look at changes to heavyrainfall events. Paradoxically, in some areas, despitethe strong summer-average drying, increases in thefrequency of heavy rainfall events are predicted.

The Hadley Centre RCM has been adapted to runon personal computers, with a domain able to beset to anywhere on the Earth, in a system calledPRECIS (Providing REgional Climate for ImpactsStudies). It is made freely available to anydeveloping country, after training provided by theHadley Centre, so that it can generate scenarios innational centres, validate them using localobservational data, and take ownership of them. Bymid-2005, eight courses in different parts of theworld had trained almost 150 modellers. Seewww.precis.org.uk for further details.

Further reading: Buonomo, E., R. Jones, C. Huntingford and J. Hannaford. The robustness of high resolutionpredictions of changes in extreme precipitation for Europe. Submitted to Quart.J.Roy.Met.Soc.(2005).

% %

44

46 Cl imate change

In 1997, Defra set up the UK Climate ImpactsProgramme (UKCIP) to encourage and facilitateadaptation to climate change. As one of the maintools to help stakeholders, in 2002 UKCIP publisheda set of detailed climate change scenarios for theUK. These were based on the Hadley Centre globalclimate model predictions, downscaled over Europeby the Hadley Centre RCM. Scenarios wereaveraged over 30-year periods centred on the2020s, 2050s and 2080s, and covered a number ofclimate quantities such as temperature,precipitation, soil moisture, relative humidity, solarradiation, sea-level pressure, wind speed, etc.

Shown in this slide is an example of the scenario forchange in seasonal average precipitation, in winterand summer, by the 2080s compared to recentclimate (1961–1990) under the Medium HighEmissions scenario. Other scenarios correspond tothe three other possible future emissions.

The scenarios replaced earlier ones published in1998, and work is well underway on the next set ofscenarios, likely to be published in about 2008.

45

Cl imate change 47

The British Isles, being at the boundary of theAtlantic Ocean and the Eurasian landmass, and atthe eastern end of Atlantic storm tracks, experiencesone of the most variable climates in the world. TheNorth Atlantic Oscillation is the main driver of winterclimate variability; in winters when the pressuredifference between the Azores and Iceland is greaterthan the average (that is, relatively lower pressureover Iceland) we tend to wetter, stormier and milderwinters; when the NAO is in the reverse phase wetend to drier, calmer but colder winters. This givesrise to a marked variability in winter rainfall.

Summer rainfall is also very variable, as shown inthe slide above of total summer rainfall over Englandand Wales simulated by the Hadley Centre modelfrom 1950–2100. (Note that the model gives areasonable representation of year-to-year variabilitybut cannot predict rainfall for a particular summer.)The downward trend in summer rainfall during the21st century due to man-made climate change isclearly seen, but it is noticeable that even in thelatter quarter of the century there are still likely tobe summers which are wetter than average, even bytoday's standards.

46

48 Cl imate change

There is substantial interest in the effects of climatechange on sea level, as the increased risk of coastalflooding could markedly affect society. Sea level willchange due to expansion of oceans as they warm,and due to the influx of water from melting ofglaciers and other snow and ice, and changes in thetwo large ice sheets in Antarctica and Greenland.Above is a plot of changes in sea level predicted bythe Hadley Centre model from the years 1860 to2100, due to each of these contributors, under aMedium-High Emissions scenario*. The dotted blueline shows the expected change of sea level due tochanges in the Greenland ice sheet. The green lineshows predicted changes due to melting of glaciersand snow on land. The red line shows the majorcomponent of sea-level rise which is thermalexpansion of ocean waters. Adding thesecomponents together gives a predicted sea-levelrise from the middle of the last century to 2100 ofabout 0.4 m, about 0.1 m of which should havealready occurred (rather less than has actually beenobserved), leaving a rise of about 0.3 m over thenext 100 years.

Changes in the Antarctic ice sheet are verydifficult to estimate, and are shown by the solidblue line above. Snowfall is predicted to increaseover Antarctica in the Hadley Centre model, andthis will act to reduce sea-level rise. If we take thisinto account, the total sea level rise may be some10 cm smaller than the estimates given above, asshown by the black line in the slide. Over the nextcentury or so, the effect of Antarctic melting mayovercome any reduction in sea level due toincreased snowfall and make Antarctica a netcontributor to sea-level rise.

Other models show different estimates for thecomponents and for the total, indicating the needto improve descriptions of all the included processes,and the inclusion of additional ones such aspermafrost melt and man-made storage in reservoirs.

* Gregory J.M. and J.A. Lowe. Projections of global and regional sea level rise using AOGCMs with and withoutflix adjustment. Geophys. Res. Lett., 27(19), 3069–3072, 2000.Further reading: IPCC TAR Chapter 11.

47

Cl imate change 49

The IPCC Third Assessment brought togetherestimates of sea-level rise from seven differentclimate models, each driven with the main SRESemissions scenarios (see earlier slide). The range ofpredicted sea-level rise was very large; taking thelowest emissions scenario with the least sensitiveclimate model gave an estimated rise of about 0.1 mover the century. On the other hand, taking thehighest emissions scenario with the most sensitiveclimate model gave a rise of almost 0.9 m. Thecentral value (not necessarily the best estimate) ofabout 0.5 m corresponds to two to four times therise in the 20th century.

At a regional scale, changes are predicted to be verydifferent from the global mean, as they dependupon regional ocean heat uptake, changes incurrents and in atmospheric pressure. All modelsshow the range of regional variation in sea-level riseis substantial, in some areas twice the globalaverage and in others almost no rise at all. However,there is little agreement between models as towhich are likely to be the most and least rapid areasof sea-level rise, although there is some agreementthat rises will be larger than average in the ArcticOcean, the US East Coast and around Japan, andsmaller in the Southern Ocean.

Further reading: IPCC TAR Chapter 11 Section 5.

48

50 Cl imate change

In order to find the future change in sea level alonga particular coastline, we must add to the rise dueto man-made climate change any effect of landmovement. The west coast of Scotland, forexample, is currently rising by about 1–2 mm peryear, as it recovers from the effect of the massiveglaciers of the last ice age (known as isostaticrebound). London, on the other hand, is currentlysinking at about 1.5 mm per year.

The main consequence of sea-level rise will probablycome from an increase in extreme high water levels,which arise from storm surges as mid-latitudedepressions or tropical storms and cyclones trackacross the area. The effect of sea-level rise on stormsurges around the British Isles was investigated aspart of the development of the UKCIP02 scenariosof climate change. This involved driving a stormsurge model developed at the ProudmanOceanographic Laboratory, Liverpool, withmeteorological predictions (pressure, winds, etc.)from the Hadley Centre regional climate model.Shown above is the change by the 2080s in theheight of a 1-in-50-year high water event, underthe Medium High Emissions scenario, with a 30 cm sea-level rise, and including the effects ofland movement.

As can be seen, increases of a metre or more inextreme high water levels, three or four times thesea-level rise, are predicted in the Thames Estuaryand southern North Sea, whereas surges aroundWales change by a similar amount to the mean sea-level rise.

The effect of sea-level rise can also be expressed asa change in the frequency of a given high waterlevel. Under the same scenario as that above,extreme high water events at Immingham, a port inthe northeast of England, which currently happenon average every 100 years, are predicted to occuras often as every seven years on average in the2080s.

49

Cl imate change 51

Greenhouse effect heating in the atmosphere israpidly transferred into surface ocean waters. It thenslowly penetrates deeper and causes more andmore of the ocean depth to expand and, hence,leads to further sea-level rise. This figure shows thesea-level rise due to ocean thermal expansion,estimated from a climate model experiment whereCO2 concentration in the atmosphere washypothetically increased by 1% per year from timezero to 70 years (that is, until it had doubled) andwas then stabilised at that concentration, that is, nofurther increase occurred. The initial blue line showsthermal expansion while the CO2 concentration wasrising, the continuing red line shows sea-level riseafter CO2 concentration had been stabilised. Despitethe fact that CO2 in the atmosphere did not changeafter year 70, the sea level carries on rising for manyhundreds of years, with only a slow decrease in therate of rise. So at any time the sea-level rise causedby the man-made greenhouse effect carries with it acommitment to an additional, inescapable, rise.

50

52 Cl imate change

The oceans of the world naturally exchange largeamounts of carbon dioxide with the atmosphere.Where oceans are warm, they outgas CO2 into theatmosphere; where they are cold they take up CO2

from the atmosphere. The global ocean is also amajor sink for man-made carbon dioxide, currentlyresponsible for absorbing about a quarter of it. Theincrease in concentration of CO2 in surface watersproduced by the absorption of man-made CO2 willreduce the chemical uptake of CO2 from theatmosphere; a reduction in this sink will tend toleave more CO2 in the atmosphere and act to speedup warming (although changes in biologicalprocesses, which also control the strength of theocean sink for CO2, may offset these chemicalchanges). Increase in ocean surface temperaturesdue to human activity will also result in less uptakeof CO2 from the atmosphere.

Basic chemistry tells us that when atmospheric CO2

is absorbed it acidifies the ocean water. The HadleyCentre coupled climate-carbon cycle model hasbeen used to predict how the surface pH of theocean will change over the period 1860–2100, andthe result is shown in this slide. The red linerepresents the global average surface ocean pH,and the blue lines either side show the spread of pHat any given time from different oceans of the worldand different seasons. The reduction in pH of about0.1 unit since pre-industrial times simulated by themodel is roughly in line with measurements. Themodel predicts a further reduction of about 0.25 bythe end of the century.

The Royal Society reported in June 2005 on theeffects of ocean acidification. Its report can befound atwww.royalsoc.ac.uk/document.asp?id=3249

Increasing ocean acidity will act to decrease CO2

uptake, in addition to decreased uptake resultingfrom increased surface CO2 concentration andtemperature. There is convincing evidence tosuggest that acidification will affect calcification, theprocess by which animals such as corals make shellsfrom calcium carbonate, and this will threatentropical, subtropical and even cold-water corals.Phytoplankton and zooplankton, which are a majorfood source for fish, may also be affected. Recentwork* concludes that key marine organisms — suchas corals and some plankton — will have difficultymaintaining their external calcium carbonateskeletons. Indications are that conditions detrimentalto high-latitude ecosystems could develop withindecades, not centuries as suggested previously.

* Orr, J.C. et al. Anthropogenic ocean acidification over the twenty-first century and its impact on calcifyingorganisms. Nature 437, 681–686 (2005).

51

Cl imate change 53

There are a growing number of centres whichdevelop and use global climate models. As part ofthe IPCC Third Assessment, the Hadley Centreorganised an intercomparison of predictions fromnine climate models, and this slide shows howglobal-average precipitation changes for each of themodels over the course of the century. In each casethe models are run with the Medium High (SRES A2)Emissions scenario, so the spread of results arisesfrom differences in the models themselves.

In the case of global-mean temperature, the modelpredictions range from 1.5 °C to 6 °C, and in thecase of precipitation, from about 1.5% to 9%.

Further reading: IPCC TAR Chapter 9 Section 3.

52

54 Cl imate change

The biggest uncertainty in climate predictions arisesbecause climate models are imperfectrepresentations of the climate system. Clouds, forexample, have a great effect on climate. Low cloudsreflect sunlight but have little effect on the escapeof infrared radiation, so they have a cooling effecton climate. High clouds, on the other hand, trapinfrared radiation but do not reflect much sunlight;they have a warming influence. The net effect in thepresent day is an overall cooling effect.

However, changes to the characteristics of clouds —their amount, height, thickness or the size of theirwater droplets or ice crystals — can drastically altertheir climate properties, and hence could changethe cooling effect into a warming one. This wouldbe a positive feedback on climate from clouds.Many other processes in the climate system willchange, and cause similar feedbacks, positive ornegative. Because different climate modelsrepresent processes in different ways, the feedbacksfrom these processes will be different, and this isthe main reason why climate models give differentpredictions for the future.

53

Cl imate change 55

We have seen earlier that models give a wide rangeof predictions in global-mean quantities; theuncertainty at a smaller scale is even bigger. If wezoom into the south-east of England, for example,we see that the IPCC TAR climate models givechanges to summer-mean rainfall by the 2080s,under Medium-High future emissions, ranging froma small increase to a 45% decrease — shownschematically in the left-hand panel. Because wehave no way of assigning the skill of each of themodels, all of the predictions must be assumed tohave the same (unknown) probability. This isobviously unhelpful to planners trying to adapt toclimate change; hydrologists deciding on whether anew reservoir should be built to avoid summer watershortages, for example. If they plan for the smallestclimate change then they could be caught out ifpredictions of greater change come about. On theother hand, if they spend large sums adapting tothe highest predictions, these may be wasted ifsmaller predictions turn out to be more realistic.

The reduction in uncertainty in predictions isunlikely to be rapid, depending as it does on hard-won improvements in our understanding of howthe climate system works. Planners therefore wishto move away from the current situation of having alarge number of different predictions of unknowncredibility, to a situation where the probability ofdifferent outcomes (for example, percentagechanges in summer rainfall) is known, as in theright-hand panel. They can then use theseprobabilistic predictions in risk assessments, todecide on the optimum adaptation strategy.

Recognising that, as explained in the last slide,models give different predictions because they usedifferent representations of the climate system, weare approaching this problem in the Hadley Centreby building large numbers of climate models, eachhaving different but plausible representations ofclimate processes; so-called ‘physics ensembles’.

54

56 Cl imate change

Each member model of the physics ensemble hasbeen used to predict the climate in a world whereCO2 is doubled, and the corresponding change in aparticular quantity (seasonal mean rainfall overEngland & Wales, for example) is taken from themodel. We then combined the results from thelarge number of models in the physics ensemble inthe form of a frequency distribution of that quantity.Finally, we gave each model a weight according toits ability to simulate current climate, and weightedthe results from all the models to give an estimateof the probability distribution of a change in therequired quantity. The results were published inNature in 2004*. The figure above shows an estimateof the probability distribution of the percentagechange in summer and winter rainfall over Englandand Wales, at a time after CO2 has doubled.

The next stages are to take these ‘doubled-CO2

world’ results and use them to predict changes overthe course of the century. We will then downscalethe results from this global model to a 50 km or 25 km scale using the Hadley Centre regional model.Because all the models in the current ensemble arevariations of the Hadley Centre climate model, wethen need to take account of predictions from otherclimate models, which may have different structuresfrom the Hadley Centre model. This process iscomplex and will take considerable effort andcomputing resources, but we aim to have initialprobabilistic predictions for use in the next set ofUKCIP climate change scenarios.

*Murphy, J.M. et al. Quantification of modelling uncertainties in a large ensemble of climate changesimulations. Nature, 430, 768–772, 2004.

55

Cl imate change 57

There are two major ice sheets which are thought tobe vulnerable to climate change: the West AntarcticIce Sheet (WAIS) which contains ice equivalent toabout 6 m of global sea-level rise, and the GreenlandIce Sheet containing just over 7 m of sea-levelequivalent. Very much larger amounts of water arelocked up in the remainder of the Antarctic icesheet, but this is not thought to be vulnerable to awarming of less than about 20 °C.

In the case of the Greenland Ice Sheet, if summerregional temperatures were to rise by about 3 °C,the ice sheet would begin to reduce in size. It wouldbe slow to disappear, perhaps half of it taking about1,000 years to melt. This critical temperature ispredicted to be reached by the end of the centuryby most combinations of climate models and futureemissions scenarios.

The West Antarctic Ice Sheet is grounded below sealevel. Its potential to collapse in response to futureclimate change is still the subject of debate andcontroversy. The IPCC TAR took the view that a lossof grounded ice leading to substantial sea-level risefrom this source was very unlikely during the 21stcentury. However, recent measurements suggestthat the melting of some ice shelves (floating seaice attached to the coastline) is leading to a speedup of glaciers and, hence, an increase in theirdischarge into the sea. The implications of this,needed before reliable predictions can be made,have yet to be understood.

56

58 Cl imate change

There have been some concerns expressed thatglobal warming could lead to massive changes inocean currents such as the Gulf Stream. Currents inthe ocean are responsible for about half the work ofthe climate system in redistributing heat betweenthe equator and the poles. The current system inthe N Atlantic is driven by ‘convection’ which takesplace in two areas, near Labrador and in theGreenland-Iceland-Norway sea. Here, the surfacewater is cooled by arctic winds, and sinks a fewthousand metres to the bottom of the ocean. Thiscool dense water then flows southwards, with aflow equivalent to a hundred Amazon rivers, crossingthe equator and heading south. The sinking coldwater in the north has the effect of drawingnorthwards warm near-surface water from the Gulfof Mexico, which travels across the North Atlantic. Itis often called the Gulf Stream, but is more properlyreferred to as the North Atlantic Drift. The heat whichit transports towards north-west Europe is part ofthe reason why countries such as the British Islesand Norway are a lot warmer than, for example,those parts of western Canada at the same latitudes.This global ocean circulation, which extends to otheroceans of the world, is known as the thermohalinecirculation (THC), as it is driven by differences intemperature and salinity of the water masses.

57

Cl imate change 59

There is strong evidence that the Gulf Stream hasswitched off more than once over the last tenthousand years, due to natural causes. We alsoknow that there are processes that have thepotential to make this happen again. Firstly, thewarming of surface waters in the convection areas,due to the man-made greenhouse effect, willreduce their density. Secondly, we expect there tobe increased rainfall over the convection areas, andthis freshwater will also act to reduce the density ofsurface waters. Thirdly, increased precipitation inhigh latitudes in a warmer world will increase theoutflow of fresh water from rivers – this has alreadybeen observed and, as mentioned earlier, recentwork at the Hadley Centre has been able toattribute this to man-made climate change. Andlastly, as the amount of sea ice decreases, a furthermechanism for driving convection (the seasonalfreezing of sea ice which rejects salt and thus makesurface waters denser) will also decrease. All thesefactors will act to reduce the density of surfacewaters and make them sink more slowly.

What would happen if the north Atlanticthermohaline circulation did switch off? The THCin the Hadley Centre climate model was artificiallyswitched off by ‘pouring’ fresh water on the oceansurface at the convection areas. The change insurface temperature in the first decade after thiswas done is shown in the slide above. The whole ofthe Northern Hemisphere cools, and around theNorth Atlantic, and particularly the Arctic Ocean,the cooling is very obvious. The UK would cool bysome 3–5 °C. The effects on extremes, for examplewinter minimum temperatures over Central England,would be very marked, and the effect of this oninfrastructure would be likely to be far worse thanthat of a gradual global warming (although nothinglike as great or as sudden as depicted in theHollywood disaster movie ‘The Day After Tomorrow’).

But this is a hypothetical experiment; do we expectthe ocean circulation to switch off as the worldwarms?

Reference: Wood, R.A., M. Vellinga and R. Thorpe. Phil. Trans. Roy. Soc. Lond. (A), 361, 1961–1975, 2003.

°C58

60 Cl imate change

The Hadley Centre climate model, which representswell the convection processes and ocean circulation,having been extensively validated against oceanobservations, has been used to predict futurechanges in the ocean circulation. The coloured linesshow the prediction of change in the North Atlanticcirculation for the four SRES emissions scenariosdescribed earlier. (The unit of current strength is asverdrup, a million cubic metres of water persecond). It is seen that all of the emissions lead to adecrease in the circulation strength of about15–25% by 2100. This will, of course, lead to acorresponding reduction in the amount of heattransported towards the UK, but this is more thanoffset by the direct greenhouse warming. As theyare derived from the same model experiment, thisreduction is taken into account in the UKCIP02climate change scenarios shown earlier. When theexperiment is continued, with greenhouse gasconcentrations stabilised in 2100, so; the Low andMedium Low scenarios show no further decline.

Although other climate models reviewed in theIPCC TAR show different rates of decline of theocean circulation, none of them shows a completeswitch-off of the Gulf Stream by 2100.

59

Cl imate change 61

In an earlier slide we saw that, averaged over the1980s, of the 7 GtC/yr emitted from human activities,about 3 GtC/yr remained in the atmosphere andabout 2 GtC/yr was absorbed by the oceans and byvegetation on land. The two ‘sinks’ together takeup about half of our emissions and moderate CO2

concentrations and reduce man-made globalwarming. However, this may not always be the casein future. As global temperatures rise, and rainfallpatterns change, we believe that several changes tocarbon absorption will take place. Firstly, in theright conditions, CO2 fertilises vegetation andspeeds up its growth; this will absorb more of ourCO2. Secondly, higher temperatures and morerainfall will encourage growth of high latitudeforests, and this will also help to mop up more ofour CO2. However, as soils get warmer, themicrobial action which breaks down humus worksfaster, and this will cause more CO2 to be emittedinto the atmosphere. Lastly, higher temperatures(and thus higher evaporation) and lower rainfall arepredicted for some forests in the tropics, and this ispredicted to cause them to die back, with theircarbon store being returned to the atmosphere.

The Hadley Centre coupled climate-carbon cyclemodel has been used* to estimate changes to theamount of carbon stored in oceans, vegetation andsoils. The figure above shows the trend in the lattertwo from 1860 to 2100. It can be seen that soils andvegetation initially act as excellent sinks of carbon,and their carbon content steadily increases.

However, as explained above, by the middle of thefirst half of the century the gradually warming soilsthen begin to emit more CO2 than they absorb, sothey act as a source rather than a sink. The samething happens with vegetation, but on a slowertimescale; there the sink-source transition is predictedto happen in the second half of the century.

The reduction in natural carbon sinks, and theireventual change into carbon sources, allows moreman-made CO2 to remain in the atmosphere, andso concentrations build-up faster — a positivefeedback between the climate and the carbon cycle.Under one business-as-usual emissions scenario,CO2 concentration in the Hadley Centre model waspredicted to rise to 750 ppm by 2100, accompaniedby a warming over land of about 5 °C. When thefeedback between climate and the carbon cycle wasincluded in the model, CO2 was predicted to rise to1,000 ppm, and global mean temperature overland to 8 °C, under the same emissions scenario.This first estimate of the strength of the feedbackhas since been repeated by other modellingcentres, which unanimously agree that climatechange will reduce the natural absorption of CO2

by the biosphere, although they find a variety ofdifferent strengths, so more research is needed toreduce uncertainties. But the potential forenhancement of global warming from this feedbackhas been clearly demonstrated.

* Cox P.M., R.A. Betts, C.D. Jones, S.A. Spall and I.J. Totterdell. Acceleration of global warming due to carboncycle feedbacks in a coupled climate model. Nature, 408, 184–187, 2000.

60

62 Cl imate change

The ultimate objective of the UN FrameworkConvention on Climate Change is to achieve“…stabilisation of greenhouse gas concentrations inthe atmosphere at a level that would preventdangerous anthropogenic interference with theclimate system.” Even after greenhouse gasconcentrations are stabilised it will take a long timefor a balance to be reached between the amount ofenergy coming in from the Sun and the amountlost to outer Space in the form of infrared radiation.If greenhouse gas concentrations were stabilisedtoday (taken as the year 2000), which (in the caseof carbon dioxide) would require a reduction inemissions of about 70%, we predict that globalaverage temperature would keep rising andeventually warm a further 1 °C above today’stemperature, as shown in the above slide, withestimates taken from a simple energy balancemodel rather than a full GCM. This is because of theinertia of the climate system, particularly the ocean,which has a large thermal capacity.

We have also used a full GCM to explore globaltemperature rise to 2100 due to two scenarioschosen for the IPCC 4AR. These follow an SRESemissions profile to 2100, after which the radiativeforcing in the model is then held constant for afurther 100 years. The temperature followingstabilisation at 2100 after the lower (B1) emissionsscenario rises a further 0.5 °C between 2100 and2200; that following stabilisation after the A1FIemissions rises almost a further 1.5 °C. There areindications that estimates of commitment using thefull GCM are less than those using a simple model.

61

Cl imate change 63

Reports referred to in the answersIPCC Intergovernmental Panel on Climate

Change, set up in 1988 under twoUN bodies

TAR Third Assessment Report from the IPCC,published in 2001

4AR Fourth Assessment Report from the IPCC,due to appear in 2007

SRES IPCC Special Report on EmissionsScenarios, published in 2000

UKCIP02 Climate Change Scenarios for the UnitedKingdom, launched by Defra/UKCIP inApril 2002

Q1: How do we know that the climate ischanging?

A1: Although several aspects of climate are changing,temperatures provide the clearest evidence. Formany decades, temperature near the surface hasbeen carefully measured at thousands of locationson land and at sea. There are a large number ofmeasurements of temperature close to the Earth’ssurface which are global in extent, from which wecan form a global average, going back to 1860.These all show temperatures higher in the past fewyears than at any time during the instrumentalperiod, even allowing for measurement uncertaintiesand gaps in the data. Global average land and seatemperatures, shown in Slide 24, show considerablevariability from year to year, but a clear underlyingtrend which shows rising temperatures until about1940, a slight downward trend from about1940–1975, and a rise of about 0.5 °C between1975 and the present day.

Three independent types of temperaturemeasurement — air temperature taken at landclimate stations and on ships at night (when theinterfering effect of solar radiation is absent) andthe temperature of the sea surface — all show (seeSlide 25) good agreement from 1900 until the lastcouple of decades, when land temperatures havebeen rising at a faster rate than sea temperatures (aspredicted to be the case for a global warming dueto increased greenhouse gases).

Temperatures have also been measured in theatmosphere; over the last 50 years or so by weatherballoons, and by satellite remote sensing since1979. In the mid-troposphere, about 5 km abovethe surface, there has been a global-mean warming— see Slide 31. Although data are sparse in tropicalregions, according to sensors on weather balloons,there seems to have been little change intemperature in the tropical mid-troposphere overthe past 25 years, which is not what modelspredict. This discrepancy and its implications arethe subject of ongoing research.

Q2: Could recent climate change be largelythe result of natural variability?

Slides 27 and 28 showed how the observations ofwarming over the past 30 years or so cannot bereplicated in the climate model if only natural factorsare included, but can be replicated once man-madefactors are added. This work used global-meanwarming, and it is possible that the good agreementwith observations could be as a result of offsettingerrors in the model, for example by the modelexaggerating the effects of man-made greenhousegases and man-made cooling aerosols.

For this reason, more detailed studies have beendone looking at the patterns of changes intemperature across the surface of the Earth andthrough the depth of the atmosphere. The resultsfrom these enabled IPCC to pronounce in the TARthat “…most of the observed warming over the past50 years is likely to have been due to the increase ingreenhouse gas concentrations”, which in turnhave been due to emissions from human activities,notably fossil fuel burning.

Since the TAR in 2001, several new studies havestrengthened this conclusion. So, whilst we cannotabsolutely exclude natural variability as the cause ofwarming over the past few decades, and it mayhave played some role, it is very unlikely that thiswill have been the sole reason. Our best estimate isthat most recent warming is due to man’s activities.

Q3: Isn’t the apparent warming due tourbanisation?

No. Slide 26 shows the results of a recent analysis atthe Hadley Centre of temperature trends over thelast 50 years deduced separately from measurementson the most windy nights and the least windynights, from the same stations. If urbanisation werecausing the observed warming, one would expectcalmer nights to have warmed more, as it is in theseconditions that the heat island effect would act towarm the station compared to its rural surroundings.In fact, Slide 26 shows no difference betweentrends on windier and calmer nights, confirmingthat urbanisation is not to blame.

Q4: Hasn’t climate variability been shown tocorrelate with solar variability?

Yes. Slide 9 shows an estimate of how solar irradiancehas changed over the last 150 years. There appearsto have been an upward trend from about 1900 to1960, but thereafter little long term change (justthe effects of the 11-year solar cycle). Simplyworking from the change in the amount of solarirradiance striking the Earth, we can calculate thatthis would have given a rise in global temperatureof about 0.1 °C; this is also shown on the slide.Even if the Sun had a much larger influence onclimate than currently thought, changes in the Suncould not explain the warming since the 1970s.

Climate change: some frequentlyasked questions, with answers

64 Cl imate change

There have been theories which seek to explainrecent global warming as due to changes in theSun’s magnetic activity (which is somewhatdifferent from the behaviour of its irradiance). Weknow that increasing solar activity tends to reducethe number of galactic cosmic rays entering theEarth’s atmosphere. Some theories argue thatgalactic cosmic rays are important for the formationof clouds, and we certainly know that clouds are animportant influence on climate. In this way, thetheory links changes in solar magnetic activity withchanges in climate. However the link betweencosmic rays and clouds is unproven, and even ifcosmic rays did cause variation in clouds, this maynot be in the right sense to explain climate change.

Q5: Aren’t greenhouse gases like methanemuch more potent than CO2, particularly if ashort time scale is considered?

Yes. The warming effect of methane and othergreenhouse gases per kilogram emitted is generallygreater that that of CO2. IPCC defines a quantitycalled Global Warming Potential which comparesthe warming effect of a greenhouse gas over agiven time period (usually taken as 100 years) withthat of CO2 (which is given a value of 1). Mostgases are more ‘potent’ than carbon dioxide,largely because the atmosphere already containsquite high concentration of CO2, and hence itsabsorption of infrared radiation (the mechanism forthe greenhouse effect) is already quite saturated. Infact, the additional infrared it traps is proportionalto the logarithm of its concentration. For othergases, such as methane, infrared absorption is stillfar from saturated, and this is the main reason whyits GWP is higher than that of CO2. The value of theGWP for a gas reflects not only its infrared absorbingcapability, but also its lifetime in the atmosphereand its density. By definition, the GWP of CO2 isunity; that of other gases is given in the textaccompanying Slide 17.

Of course, the warming effect of a particular gas willbe a combination of its GWP and its total emissions,and because man-made emissions of CO2 are muchgreater than any other gas, its warming effect willbe greater, despite its low GWP. This is illustrated inSlide 17, where roughly two-thirds of the man-madewarming effect over the next 100 years are projectedto be due to CO2 emissions.

Q6: Surely the oceans and the land surfaceabsorb large amounts of man-made CO2?

Yes they do, although it is by no means certain thatthey will continue to absorb as much as they donow. We estimate for the decade of the 1980s (seeSlide 11) that fossil-fuel burning injected on averageabout 5.4 GtC/yr (billion tons of carbon, in theform of carbon dioxide, per year) into theatmosphere. In addition, changes to land use,mainly deforestation, added another 1.7 GtC/yr. Weobserve that atmospheric concentrations of CO2 arerising at about 2 ppm per year, which equates to anincrease in the burden of carbon of some 3.3GtC/yr. Of the remaining 3.8 GtC/yr, we estimatethat oceans absorb about 1.9 GtC/yr and vegetationand soils a further 1.9 GtC/yr. Thus, the ocean andland provide a free ‘buffering’ service by absorbing

about half of the carbon we emit. However, thismay not always be the case in future. As globaltemperatures rise, and rainfall and temperaturepatterns change, we believe that several changes tocarbon absorption will take place, as described inSlide 60. The net result of these changes is predictedto be a reduction in the strength of the carbon sinkin vegetation and soils, leaving more of the carbondioxide that we emit left in the atmosphere, and amore rapid increase in CO2 concentration andtemperature rise than without this reduction in sinks.Although there is agreement amongst modellers thatclimate change will reduce the natural absorptionof CO2 by the biosphere, different models estimatedifferent reductions. But the potential forenhancement of global warming from this feedbackhas been clearly demonstrated.

Q7: As natural emissions of carbon dioxide arevery much greater than those from humanactivities, surely the effect of man isinsignificant?

The exchange of ‘man-made’ carbon dioxide betweenman-made emissions, atmosphere, ocean and land,is about 7 GtC per year, as shown in Slide 12, whichalso shows much larger natural exchanges betweenatmosphere and ocean (about 90 GtC/yr) andatmosphere and land (about 60 GtC/yr). However,these natural exchanges have been in balance formany thousands of years, leading to the pre-industrialconcentration of CO2 remaining steady at about280 ppm (see Slide 14). The effect of the additionalman-made emissions is to unbalance the budgetand lead to the rise in concentrations seen sinceabout 1850, also shown on Slide 13.

Q8: Will aerosols help reduce climate changeor even result in cooling?

There are many different types of aerosols — smallparticulates — in the atmosphere which are affectedby human activity. Some, such as black carbon(soot), are emitted directly from man-madeprocesses, and some are generated from otherman-made emissions, such as the sulphate aerosolswhich are formed in the atmosphere from sulphurdioxide emissions from power stations, transport,etc. Some, such as mineral dust from deserts, areentirely natural, but their concentration in theatmosphere (and hence their effect on climate)could be changed if man-made climate changeleads to desertification.

As can be seen from Slide 20, some aerosols, suchas sulphates, predominately act to reflect back solarradiation (both directly and indirectly, see Slide 18)and hence exert a cooling influence on climate.Others, such as black carbon, absorb solar radiationand have a warming effect. The slide also shows howvery uncertain the warming and cooling estimatesof these aerosols is, compared to the relative certaintyof the effect of greenhouse gases. Nevertheless, it isvery unlikely that there has been sufficient aerosolcooling to have offset the warming effect of man-made greenhouse gases, and in the future, asconcentrations of cooling sulphate aerosols arelikely to decline, their lessening cooling effect mayhave the consequence of accelerating warming.Indeed, recent Hadley Centre estimates are that the

Cl imate change 65

exceptionally hot 2003 summer over continentalEurope would happen typically more frequentlythan every decade in the absence of cooling fromsulphate aerosols.

Q9: What is global dimming and whatrelevance does it have to climate change?

‘Global dimming’ is the term used to describe theobservations from surface instruments showing ageneral reduction in the amount of solar radiationreaching the ground since about 1960, globallyamounting to 2–3% per decade, up to about 1990.The dimming is variable from place to place, withsome sites even showing a brightening over theperiod, but greatest in northern mid-latitudes.However, more recent research indicates that thistrend reversed in about 1990 and since then therehas been some ‘global brightening’, although beingindirectly measured from satellites these morerecent estimates may be less robust. It seems likelythat the reductions, and perhaps the recentincreases, may be due to changes in aerosols suchas sulphates and soot (black carbon). The mostrecent version of the Hadley Centre climate model(HadGEM1), which includes both sulphate aerosolsand soot, does simulate a reduction in surface solarradiation, though not as great as that actuallyobserved. Neither the observations nor theimplications for predictions of climate change areyet clear, and this is a subject of active research.

Q10: How reliable are climate models?

Climate models are a mathematical description of theprocesses in the Earth’s climate system; atmosphere,ocean, land, cryosphere. The representation ofclimate processes in the model are based onexperimental measurements in the real atmosphere,ocean etc, and these can be chosen within theconstraints of these experiments to give the bestpossible agreement between model simulation ofcurrent climate and observations. We evaluate theirreliability in a number of ways. Firstly by comparingtheir representation of the current climate andobservations, including not just means but variabilityand extremes. Secondly, by driving them with thebest estimates of changes to climate forcings overthe last 150 years (natural, such as volcanoes andsolar radiation, and man-made such as greenhousegases and aerosols) and comparing the simulationof climate change from the model (sometimes calleda ‘hindcast’) with observations of trends (in, forexample, global mean temperature) over the sameperiod. This is shown in Slide 28. Lastly, somevalidation can be carried out by comparing modelsimulation of climates many thousands of years agowith reconstructions of climate of the period (socalled palaeoclimatologies). Validation exercises suchas these provide compelling evidence that, at leastin terms of gross temperature response, the modelis effectively reproducing what has been observed,and this gives us confidence that the models areadequate tools for the prediction of future climates,albeit with the sort of uncertainty described in Slide52 (at a global mean scale) and in Section 3.5 ofthe UKCIP02 report (at the scale of the UK). Moredetail on model validation and performance can befound in Chapter 8 (Model Evaluation) of IPCC TAR.

Q11: Weather forecasts aren’t accurate formore than a few days ahead, so how can wepossibly predict climate over the next 100years?

Although they are made by the same sort ofmathematical model, weather forecasts andclimate predictions are really quite different. Aweather forecast tells us what the weather (forexample, temperature or rainfall) is going to be ata certain place and time over the next few days; itmight say, for example, that there will be a bandof heavy rain moving across Somerset tomorrowmid-morning.

A climate prediction tells us about changes in theaverage climate, its variability and extremes. Forexample, it might say that the average temperatureof summers in Somerset in 40–60 years time will be4 degrees higher than it is currently, it will enjoy onaverage 25% more rain in winter with three timesthe current number of heavy rainfall events, and50% less rain in summer. It will not make a specificforecast such as: it will be raining in Somerset onthe morning of 15 October 2044.

Q12: Shouldn’t we wait until the predictionsare more certain before taking action tocontrol or prepare for climate change?

This is not a scientific question and, hence, outsidethe remit of this booklet. However, there are acouple of scientific points that can be made whichare very relevant to the policy process. The first isthat reducing the uncertainty in predictions takestime, because it requires (a) a better understandingof the processes in the atmosphere, ocean,cryosphere and on land, which control climate,and (b) improvements to models to reflect this newunderstanding and to run at a higher resolution,which in turn demands greater computer capacity.Progress is being made on both these fronts, forexample the resolution of the newest HadleyCentre model (HadGEM1) is eight times greaterthan the previous model, HadCM3, and includesnew processes such as the effect of soot and theridging of sea ice. Nevertheless, benefits fromdetailed research into climate processes whichinforms the models, and the development of themodels themselves, take time to feed through intoimproved predictions.

The second point is that, because of the largeinertia of the climate system, at any given time pastgreenhouse gas emissions will have warmed theworld but also have committed us to a furtherwarming even if no further greenhouse gases wereto be emitted. The longer we leave any decisionand action to control climate change, the greaterthe commitment to future climate change (and itsimpacts) that will have built up.

Q13: Will ice sheets melt with climate change?

The two major ice sheets are on Greenland and inthe Antarctic. The Greenland Ice Sheet containsenough water to contribute about 7 m to sea level,and the West Antarctic ice sheet (WAIS), which isthe part of the Antarctic ice sheet most vulnerableto climate change, contains about 6 m.

66 Cl imate change

A sustained rise in local temperatures of about 3 °C,equivalent to a global-mean warming of about 1.5 °C, which is likely to be reached by the end of thecentury if man-made emissions are not controlled,would melt the Greenland Ice Sheet, although it isestimated that this would take a few thousand years.A major collapse of the WAIS is thought to be veryunlikely during the 21st century, although recentmeasurements suggest that contributions to sea-level rise from this source may be greater thanpreviously estimated.

Q14: Isn’t the evidence of temperature changein the Arctic and Antarctic inconclusive?

Although there is a clear rising trend in globallyaverage temperature, there are large variations intrend from region to region. This is a consequenceof natural variability of climate, which gets larger aswe focus on smaller and smaller areas. This resultsin some areas warming less than the global average— or even cooling — and some areas warmingmore. In addition, naturally variability tends to begreater at high latitudes. These two factors lead tothe Arctic and Antarctic having a wide variety oftemperature changes. For example, the AntarcticPeninsula has warmed dramatically over the past 50years, whereas at the same time some inland areasof east Antarctica have cooled. However, recentresearch suggests that changes to the winds overAntarctica, which may have been brought about bystratospheric ozone depletion, have played asignificant role in the peninsular warming and thecontinental interior cooling.

Q15: How likely is the Gulf Stream to stopflowing? Will this make Europe colder?

The Gulf Stream (or North Atlantic Drift, to give itits proper title) brings warmer water from lowerlatitudes to the north-east Atlantic, and gives NW Europe a milder climate than it wouldotherwise have.

The mechanism driving circulation in the N Atlantic,of which the Gulf Stream is a part, is shown in Slide57. This mechanism could be affected by man-madeglobal warming in several ways, for example byincreased rainfall over the N Atlantic, and hencethere is the potential for the Gulf Stream to bereduced, or even switched-off, by man’s activities.

When we use the Hadley Centre climate model tolook at the response of the N Atlantic oceancirculation to future man-made emissions, shown inSlide 59, we see that reductions of about 20% by2100 are predicted, rather than a completeshutdown. Other good climate models see greateror lesser reductions, but none produces a shutdownover the next 100 years.

The Hadley Centre model has also been used toinvestigate the impact on climate of a hypotheticalshut-down of the THC. It predicts that whole ofthe northern hemisphere would be cooled,especially the N Atlantic; the UK might see acooling of 3–5 °C. Daily minimum temperatures incentral England in winter could plunge by 10 or20 °C, and this would likely have a bigger effecton UK society than global warming. However, as

was pointed out above, this is a ‘what-if’ scenarioand not a prediction.

The model predictions of only partial shut-down ofthe THC seem reassuring, but we do not fullyunderstand the reasons for the stability of the oceancirculation, and there have been recentmeasurements in the N Atlantic which seem to beat variance with model simulations. Hence, researchcontinues to quantify the risk of this potentiallyhigh-impact outcome of climate change.

Q16: Isn’t another ice age due soon? Andwon’t it counteract global warming?

Over the past half million years or more, theworld has alternated between ice ages andinterglacials (periods between ice ages), withinterglacials occurring every 100,000 years or so.We have been in the present interglacial for about10,000 years. Evidence is strong for this behaviourto be due to changes in the Earth’s orbit aroundthe Sun, and the angle of its rotational axis,usually referred to together as ‘astronomicalforcing of climate’. This theory was formalised byMilankovic in the 1920s, and has been wellconfirmed by records from ice cores, oceansediments, etc. Thanks to our knowledge oforbital mechanics these astronomical changes canbe predicted, and it appears that astronomicalforcing will be of little significance over the next40,000 years or so, so the next ice age will be avery long time hence. Thus is it on a verydifferent timescale to man-made global warmingand cannot counteract it; if no action is taken tolimit fossil-fuel emissions, for instance, climatewill have changed very substantially by the timethe next ice age starts.

Q17: Will natural methane emissions enhanceman made emissions?

Substantial quantities of methane are emittednaturally from wetlands, and this emission isexpected to change as wetlands change. Changingrainfall patterns will cause some wetland areas toincrease in extent, others to decrease, and increasesin temperature will act to increase emissions fromwetlands. One version of the Hadley Centre climatemodel includes a description of wetland methane,and this predicts an increase in natural wetlandemissions by the end of the century equivalent tothe amount of man-made emissions projected forthat time, thus leading to a more rapid rise inmethane concentrations, and hence warming.

On the other hand, the chemical reactions in theatmosphere which destroy methane are expected tobecome more efficient in future, largely as a resultof increased water vapour. This will act as anegative feedback on methane amounts.

Methane is also stored in permafrost, and it is likelythat some of this will be released as surface warmingextends into the permafrost and begins to melt it.

Finally, huge amounts of methane are locked up inmethane hydrates (methane clathrates) in the oceans.They are currently at high enough pressures andtemperatures to make them very stable. However,

Cl imate change 67

penetration of greenhouse effect heating into theoceans may destabilise them and allow some of themethane to escape into the atmosphere. Thepotential for this to happen is very poorly understood.There is concern that this may be another positivefeedback not yet included in models, although thereis little evidence for this from the behaviour ofmethane during the large temperature swingsbetween ice ages and interglacials, and in particularover the last 50,000 years.

Q18: Are ozone depletion and climate changepart of the same thing?

Not really, although there are links between thetwo. The depletion of ozone in the stratosphereover Antarctica (the ‘ozone hole’) was firstdiscovered by scientists from the British AntarcticSurvey in the mid-1980s. It is caused mainly byemissions of man-made chlorofluorocarbons(CFCs), which find their way into the stratospherewhere they decompose into chlorine compoundswhich destroy ozone each autumn. Despite the factthat emissions of CFCs have been very severely cutback by the Montreal Protocol, because they have alifetime of order 100 years, their concentration inthe atmosphere has only recently started to turndown, and the ozone hole is expected to remain aslarge as it is now for decades to come, before itslowly recovers.

Links with climate change are threefold. Firstly, theCFCs which deplete ozone, and also some of theirozone-friendly replacements, are greenhouse gasesand so also contribute directly to global warming.Secondly, the reduction in stratospheric ozone,both over Antarctica and more generally globally,acts to cool climate slightly; see Slide 20. Lastly,there is concern that increasing concentrations ofCO2 from fossil-fuel burning, because it is coolingthe stratosphere (see Slide 31), aids the formation ofthe small particles in the stratosphere on whichchemical reactions take place, and may beprolonging the ozone hole.

Q19: What will the impacts be of man-madeglobal warming?

The impacts of climate change on society andeconomies will be many and various, in sectorssuch as agriculture, water resources, ecosystems,health, coastal communities, etc. It is too broad atopic to be covered in this Q&A section, but iscomprehensively addressed in the report fromWorking Group 2 of the IPCC TAR, which alsocontains a shorter Technical Summary andSummary for Policymakers.

Recent UK research on the global impacts can beseen in the April 2004 special edition of thejournal Global Environmental Change, edited byM. Parry.

Q20: Will increased CO2 in the atmospherestimulate plant growth?

Plant growth depends upon several factors. Plantsrequire sufficient warmth, moisture, light andnutrients in the soil to photosynthesise, that is, todraw down CO2 from the atmosphere into the body

of the plant. If these other environmental factors areadequate then higher concentrations of CO2 in theatmosphere will indeed enable plants to grow morerapidly. However increasing CO2 concentration alsochanges climate, and if this becomes too warm ortoo dry then plants will no longer be able to takeadvantage of the CO2 fertilisation effect. Hencethere is a balance; over the past century theenhanced growth has dominated and vegetationacross the globe has acted as a vital sink for man-made CO2 emissions. But, as described in Slide 60,our research indicates that in future the beneficialeffect of higher CO2 concentrations will be reducedas the associated climate change in some areas willreduce the ability of vegetation to absorb man-made CO2.

A further concern is the sensitivity of plant growthto concentrations of ozone, which is expected toincrease in the lower atmosphere due to reactionsbetween man-made emissions such as nitrogenoxides and hydrocarbons. Ozone can have adamaging effect on plant stomata and so there isa risk of reduction in vegetation productivity asozone increases.

Q21: What do the different emissionsscenarios driving the climate changepredictions represent?

Future emissions of greenhouse gases from humanactivities will depend upon factors such aspopulation growth, economic development, energyuse, technological change, society’s attitudes andpolitical leadership. Obviously, we cannot knowhow all these factors will change, and whatpathways emissions will follow in the future, but wecan generate possible scenarios; theIntergovernmental Panel on Climate Change didthis in its Special Report on Emissions Scenarios(SRES) in 2000. It considered various ‘storylines’ ofhow the world will develop and used models toestimate emissions which would follow from thesestorylines. All of the emissions scenarios are ‘non-interventionist’; that is, they assume no policies toreduce emissions for the purpose of mitigatingclimate change.

Scenarios of climate change over the UK werepublished in 2002 for Defra and the UK ClimateImpacts Programme (UKCIP02). They were basedon climate predictions from Hadley Centre models,down to a resolution of 50 km. The headlinepredictions were for warming throughout the yearbut particularly in summer; less rain in summer butmore rain in winter; greater frequency of heavyrainfall events in winter; reductions in snowfall andfrosts; continued rise in sea level; increase in thefrequency or height of coastal high water events.There is considerable regional variation in thechanges; in broad terms they are expected to begreatest in the south and east, smallest in the northand west. The scenarios report stressed theuncertainty in the scenarios and suggested someways of handling this. More detail is available fromwww.ukcip.org.uk

68 Cl imate change

Q22: Isn’t climate change going to be a goodthing in the UK?

It may indeed be more pleasant to have warmertemperatures in autumn, winter and spring.However, summers in the UK, especially in thesouth-east and in cities, could be uncomfortablywarm, leading to heat-related medical problemsand aggravating respiratory conditions. There is awell known link between high temperatures andmortality rates. The exceptionally hot summer of2003 resulted in 22–35,000 additional heat-relateddeaths across the continent of Europe, and some¤10 billion uninsured crop losses. On the otherhand, less cold conditions in winter would lead tofewer deaths from hypothermia.

Of course, climate change means much more thansimply an increase in temperatures. The summerswill probably become drier as well as hotter,leading to an increased risk of drought and pressureon water resources. Winters are likely to bring heavyrainfall events more frequently, with increased risksof urban and river flooding. As sea level rises, ourcoastline, especially in the south and east will beincreasingly at risk, and more frequent high waterevents would be particularly damaging if the levelof protection is not raised.

Q23: How will climate change impact on ourlives in the UK?

Climate change will have impacts not only on theenvironment, but also on society and the economy.To find out more about these, please contact theUK Climate Impacts Programme, which is based atthe University of Oxford. Visit www.ukcip.org.uk,for further information.

Q24: Which of the UKCIP02 climate scenariosis most likely?

The four climate change scenarios developed forthe UK are based on four possible future pathwaysof man-made greenhouse gas emissions, derivedfrom the IPCC SRES report, as shown in Slide 38 inthe case of CO2. IPCC states that these should notbe regarded as equally probable, but there is noinformation on the relative likelihood of each. Someorganisations are attempting to develop probabilisticemissions scenarios, but these are not yet at thepoint where they are reliable enough to be used asthe basis for climate change scenarios. In the case ofthe UK climate change scenarios, it is best to considerthe full range, rather than trying to identify onescenario as the most probable.

Cl imate change 69

David AcremanPaul AgnewRob AllanTara Ansell

Maria AthanassiadouHelen Banks

Rosa BarcielaBernd BeckerMike Bell

Nicolas BellouinMartin Best

Richard BettsBalakrishnan Bhaskaran

Alejandro Bodas-SalcedoPenny Boorman

Ben BoothOlivier BoucherPhilip BrohanAnca BrookshawKate Brown

Simon BrownErasmo BuonomoEleanor Burke

Neal ButchartJohn CaesarMick Carter

Helen ChampionNikos ChristidisRobin ClarkAdam ClaytonHolly Coleman

Bill CollinsMat Collins

Andrew ColmanDave Croxall

Michel CrucifixIan Culverwell

Stephen CusackMike DaveyChris DeardenCraig DonlonChris Durman

Ian EdmondPete Falloon

David FeredayChris FollandGary Fullerton

Nic GedneyPip Gilbert

John GlosterPeter GoodChris Gordon

Margaret GordonRichard Graham

Christina Greeves Jonathan Gregory

Dave GriggsJim Gunson

Mark HackettJen Hardwick

Glen HarrisChris HarrisMark Harrison

Robert HarrisonDavid Hassell

Jim HaywoodDavid Hein

Debbie HemmingFiona HewerChris Hewitt

Emma HiblingJulian Hill

Richard HillAdrian Hines

Tim HintonMartin Holt

Matthew HortTom HowardMatt Huddleston

Pat HyderSarah Ineson Bruce Ingleby

William IngramPaul James

Geoff JenkinsVenkata Jogireddy

Tim JohnsColin JohnsonAndy Jones

Andrew JonesRichard Jones

Chris JonesGareth JonesManoj Joshi

Ann KeenJohn Kennedy

Jamie KettleboroughJohn King

Jeff KnightJian-Guo LiSpencer Liddicoat

Irina Linova-Pavlova Linda LivingstonJason Lowe

Alistair ManningMatthew Martin

Gill Martin Stuart MayhewMark McCarthyMike McCullochRuth McDonald

Alison McLarenPeter McLean

Doug MiddletonJohn Mitchell

Wilfran Moufouma-Okia Steve MullerworthJames Murphy

Mo MylneFiona O’ConnorEnda O’DeaMatt PalmerJean Palutikof

Anne PardaensDavid ParkerVicky Pope

Norah PritchardJamie RaeNick Rayner

Alison Redington

Jeff Ridley Mark Ringer

Malcolm RobertsGuy RobinsonJosé RodriguezBill Roseblade

Dave RowellMaria Russo

Derrick RyallMike SandersonAndy Saulter

Michael SaunbyAdam Scaife

Yvonne SearlAlistair Sellar

Cath Senior David SextonJohn Siddorn

Stephen SitchDoug SmithFiona SmithJohn Stark

Sheila StarkDave StorkeyPeter Stott

Simon TettDavid ThomsonPeter Thorne

Thomas ToniazzoIan Totterdell

Allyn TreshanskyPaul van der Linden

Michael VellingaMark Webb

Helen WebsterPaul Whitfield

Keith Williams Simon WilsonClaire Witham

Richard WoodTamasine Woods Stephanie Woodward

Peili Wu

Met Office Hadley Centre Staff, 19 September 2005

Met Office FitzRoy Road Exeter EX1 3PB United KingdomTel: +44 (0)1392 885680 Fax: +44 (0)1392 885681 Email: enquiries@metoffice.gov.uk www.metoffice.gov.uk

Produced by the Met Office © Crown copyright 2005 05/0221 Met Office and the Met Office logo are registered trademarksPrinted on Revive Uncoated 80% post-consumer waste paper