the global climate system review 2003 - vsamp · of the global climate system review provides a...

socio-economic development – environmental protection – water resources managementsocio-economic development – environmental protection – water resources management

The Global Climate System Review 2003

WMO-No. 984

The Global Climate System Review 2003

WMO-No. 984

2005

Front cover:Europe experienced a historic heat wave during the summer of 2003. Compared to the long-term climatological mean,temperatures in July 2003 were sizzling. The image shows the differences in daytime land surface temperatures of 2003to the ones collected in 2000, 2001, 2002 and 2004 by the moderate imaging spectroradiometer (MODIS) onNASA’s Terra satellite. (NASA image courtesy of Reto Stöckli and Robert Simmon, NASA Earth Observatory)

Reference: This publication was adapted, with permission, from the “State of the Climate for 2003”, published in the

Bulletin of the American Meteorological Society, Volume 85, Number 6, June 2004, S1-S72.

WMO-No. 984© 2005, World Meteorological OrganizationISBN 92-63-10984-2

NOTE

The designations employed and the presentation of material in this publication do not implythe expression of any opinion whatsoever on the part of the Secretariat of the WorldMeteorological Organization concerning the legal status of any country, territory, city orarea, or of its authorities, or concerning the delimitation of its frontiers or boundaries.

3

ContentsPage

Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Chapter 1: Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81.1 Major climate anomalies and episodic events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81.2 Chapter 2: Global climate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91.3 Chapter 3: Trends in trace gases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91.4 Chapter 4: The tropics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101.5 Chapter 5: Polar climate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101.6 Chapter 6: Regional climate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Chapter 2: Global climate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122.1 Global surface temperatures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122.2 Upper-air temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132.2.1 Lower troposphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132.2.2 Mid-troposphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142.2.3 Lower troposphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142.3 Global precipitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142.4 Northern hemisphere snow cover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Chapter 3: Trends in trace gases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183.1 Carbon dioxide (CO2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183.2 Methane (CH4) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193.3 Carbon monoxide (CO) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203.4 Decreases in ozone-depleting gases in the troposphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Chapter 4: The tropics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224.1 ENSO and the tropical Pacific . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224.1.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224.1.2 Equatorial Pacific Ocean sea-surface and sub-surface temperature evolution . . . . . . . . . . . . . 224.1.3 Atmospheric circulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234.2 Tropical storms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244.2.1 Atlantic hurricane season . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244.2.2 Pacific tropical storms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254.2.3 South-west Indian Ocean tropical storms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

Chapter 5: Polar climate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305.1 Antarctic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305.1.1 Surface climate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305.1.2 Antarctic stratospheric ozone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315.2 The Arctic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

Chapter 6: Regional climate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346.1 North America . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346.1.1 Canada . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

4

Page

6.1.2 United States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356.1.3 Mexico . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 376.2 Central America . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 386.3 South America . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 396.4 Europe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 406.4.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 406.4.2 Summer heat wave . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 416.5 Africa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 436.5.1 North Africa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 436.5.2 West Africa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 446.5.3 Eastern Africa/the Greater Horn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 456.5.4 Southern Africa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 476.6 Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 486.6.1 China . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 486.6.2 South-west Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 486.6.3 Russia (including European Russia) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 496.6.4 Monsoon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 506.6.5 East Asian monsoon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 516.7 Australasia and the South-West Pacific . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 526.7.1 Australia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 526.7.2 South-West Pacific . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 546.7.3 New Zealand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

Chapter 7: Seasonal summaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

5

Lisa V. Alexander, David E. Parker and TaraAnsell (Hadley Centre for ClimatePrediction and Research, Met Office,Bracknell, Berkshire, UK)

Peter Ambenje (IGAD Climate Predictionsand Applications Centre, Nairobi,Kenya)

Tercio Ambrizzi (University of Sao Paulo,Brazil)

Omar Baddour (National MeteorologyDirection, Rabat, Morocco)

Gennady Belchansky (Institute of Ecology,Moscow, Russia)

Gerald D. Bell, Muthuvel Chelliah, TimothyEichler, Michael S. Halpert, Kingtse Mo,and Wassila M. Thiaw(NOAA/NWS/NCEP Climate PredictionCenter, Washington D.C., USA)

Michael A. Bell, Suzana J. Camargo andEmily K. Grover-Kopec (InternationalResearch Institute for ClimatePrediction, New York, USA)

Eric Blake and Richard Pasch (NOAA/NWSNational Hurricane Center, Miami,Florida, USA)

Olga N. Bulygina (All-Russian ResearchInstitute of HydrometeorologicalInformation, Obninsk, Russia)

Philippe Caroff (Tropical Cyclone WarningCentre, RSMC La Réunion, Météo-France)

John C. Christy (University of Alabama,Huntsville, Alabama, USA)

Miguel Cortez Vázquez (ServicioMeteorológico Nacional, Mexico City,Mexico)

Arthur Douglas (Creighton University,Omaha, Nebraska, USA)

David C. Douglas (USGS Alaska ScienceCenter, Juneau, Alaska, USA)

Sheldon Drobot (The National Academies,Washington, D.C, USA)

Karin L. Gleason, Jay H. Lawrimore, DavidH. Levinson, Matthew Menne, Anne M.Waple (STG, Inc.), Connie Woodhouse

and David Levinson (NOAA/NESDISNational Climatic Data Center, NorthCarolina, USA)

Brad Garanganga (Drought MonitoringCentre, Harare, Zimbabwe)

Stanley Goldenberg and ChristopherLandsea (NOAA/OAR/AOML HurricaneResearch Division, Florida, USA)

James Hansen (NASA Goddard Institute forSpace Studies, Columbia University,New York, USA)

Malcolm Haylock (Climatic Research Unit,University of East Anglia, UK)

Rupa Kumar Kolli (Indian Institute forTropical Meteorology, Pune, India)

Charlotte Mcbride (South African WeatherService, Pretoria, South Africa)

Simon Mcgree (Fiji Meteorological Service,Fiji)

David Phillips (Environment Canada,Ottawa, Ontario, Canada)

David A. Robinson (Rutgers University,New Jersey, USA)

Fatimeh Rahimzadeh (Islamic Republic ofIran Meteorological Organization,Tehran, Islamic Republic of Iran)

Jim Salinger (National Institute of Water andAtmospheric Research, Newmarket,Auckland, New Zealand)

Russell C. Schnell and Robert S. Stone(NOAA/ERL Climate Monitoring andDiagnostics Laboratory, Boulder,Colorado, USA)

Andrew Watkins (Australian Bureau ofMeteorology, National Climate Centre,Melbourne, Australia)

Panmao Zhai (China MeteorologicalAdministration, Beijing, China)

Authors

6

Contributors

Scott Stephens, Candace Tankersley, TrevorWallis, Richard Heim and Stuart Hinson(National Climatic Data Center, USA), NickRayner (Hadley Centre for ClimatePrediction and Research, UK), GarethMarshall (British Antarctic Survey, UK), BillChapman (University of Illinois atUrbana–Champaign, USA), C.D. Keeling,K. Thoning, S.A. Montzka, J.H. Butler,T. Thompson, D. Mondeel, J. Elkins,P.C. Novelli and E.J. Dlugokencky(NOAA/Climate Monitoring and DiagnosticsLaboratory/OAR, USA), Thomas Estilow(Rutgers University, USA)

Acknowledgements

Support was provided through a grant fromthe NOAA Office of GlobalPrograms/Climate Change Data andDetection Program and the WMO/ICSU/IOCWorld Climate Research Programme.

7

Since 1993, the World MeteorologicalOrganization (WMO), in cooperation withits Members and through its World ClimateProgramme Department and theCommission for Climatology, has issuedannual Statements on the Status of theGlobal Climate. The 2003 Statementdescribed extreme weather events andprovided a historical perspective on someof the variability and trends that hadoccurred during that year. The 2003 editionof the Global Climate System Reviewprovides a wider geographical perspectiveand complements the Statement. Both ofthese publications provide input to theperiodic assessments made by theWMO/United Nations EnvironmentProgramme (UNEP) IntergovernmentalPanel on Climate Change (IPCC).

The information contained in thisReview enhances the scientificunderstanding of the changes in climateand the associated impacts that haveoccurred in the past, making it possible toimprove on our projections for the future.Through continuing research and thecollection of consistent and completeobservations by WMO and its Members,progress towards an even betterunderstanding of the Earth’s climate systemis possible.

Some of the influence of weather andclimate on human well-being and on theenvironment are highlighted in the Review.Tropical cyclones in various parts of theworld resulted in loss of life anddestruction of property. Droughts affectedthe livelihood of many people, and heatwaves caused thousands of deaths inEurope and South-West Asia. On the otherhand, climate variability also producedbenefits to society—from abundance ofenriching sunlight for vegetable andorchard crops in western Europe to abovenormal rainfall across the Sahelian region

of western Africa and enhancedprecipitation over Afghanistan andneighbouring countries.

One of the major challenges for themeteorological and hydrologicalcommunities is the need to contribute toimproved protection of life and property.Enhanced weather, climate and hydrologicalservices are being implemented tocontribute to reducing adverse human,social and economic impacts of naturaldisasters and of extreme weather andclimate events. Among others, this is donethrough increased awareness andpreparedness of people and societies toface such events.

Through its Programmes, WMOcontributes actively to the timely provisionof authoritative climate statements, reviews,assessments and descriptions which makeimportant contributions to sustainabledevelopment and to the implementation ofthe United Nations MillenniumDevelopment Goals.

This Review is produced under theauspices of the Climate System MonitoringProject of the World Climate Data andMonitoring Programme (WCDMP), initiatedin 1984 following a recommendation of theNinth World Meteorological Congress(1983).

Special appreciation is extended to DrAndrew Watkins (Australia) for his excellentwork as coordinator of this edition of theReview.

(M. Jarraud)Secretary-General

Foreword

8

1.1 Major climate anomalies and episodic events

The year 2003 brought a wide range ofunusual weather and climate-related eventsto all corners of the globe. This in itself isnot unusual—every year brings with it asuite of anomalies. However, 2003 wasnotable for a number of events, whichappeared to be extraordinary, even giventhe background of a decaying El Niñoevent in the tropical Pacific.

In the South Pacific, two tropicalcyclones impacted upon the region duringJune, the Austral winter, while for thesecond year in a row, a south-west IndianOcean tropical cyclone made landfallduring May. In Australia, massive bushfiresfollowed 11 months of arguably (givenrecord high temperatures) its severestdrought on record. In Canada, the worstwildfire season on record struck BritishColumbia, while in the United States its

most costly fires ever burnt a large tract ofsouthern California. In the United StatesMidwest, 546 tornadoes were reportedduring May, exceeding the previous recordfor any month by 145. In the Arctic, thesea-ice annual minimum extent was thesecond lowest on record (after 2002).Europe experienced a massive andunprecedented summer heat wave, whichcontributed to many thousands of deathsand resulted in the warmest summer onrecord for France, Germany, Spain andSwitzerland. Conversely, record coldtemperatures and anomalous June snowfallsstruck European Russia. Finally, inAntarctica the ozone hole was the largeston record at 28 million km2. A summary ofthese and other global climate events isillustrated in Figure 1.1.

These events all occurred to abackground of global mean temperature+0.46°C above the 1961-1990 annualaverage, making 2003 the third warmestyear in the instrumental record since 1861,just behind the previous year, 2002

Chapter 1

Executive Summary

Western U.S.Western U.S.Continuation of multi-year drought conditions

CanadaAnnualanomalies +1-2°C

Eastern U.S.Persistent wetness

EuropeMajor heatwave in summer.21 000 deaths

Hurricane IsabelCategory 5 in Atlantic. Category 2 at landfall.16 deaths. Damageestimated at US$ 2.3 billion

Hurricane JuanCategory 2 atlandfall. Easterneyewall windsdirectly overHalifax. 8 deaths

Atlantic hurricane season Above average activity.16 named storms.7 hurricanes

Western wildfiresCostliest. Over 300 000hectares burned inCalifornia in October

Moderate El Niño fades toneutral conditions bynorthern hemispherespring

AlaskaAbove average warmth in all 4 seasons

Annual anomalies of +1-2°C.Dry in central andwestern Europe

Western AsiaaAsiaAsiaarougLong-term dLong-term drougroug

v

Dry in eastern Australia

Zimbabwe and MozambiqueDrought early in the year. Relief in austral spring

Ethiopia, EritreaDrought continued to affectfood supplies

Indonesia, Malaysiaand the PhilippinesHeavy monsoon rainfall inDecember caused floodingand landslides

Tropical Storm Linfa - Crossed the Philippinesin May with wind speeds of 100 km/hrTyphoon Soudelor - Peak wind speeds of215 km/hr in June. Heavy rain in JapanTyphoon Imbudo - Strongest typhoon to hitthe Philippines in 5 years with wind speeds of240 km/hr. Also impacted China in JulyTyphoon Etau - 165 km/hr winds in Augustand over 400 mm rain in JapanTyphoon Krovanh - 176 km/hr peak winds inChina. Also impacted Viet Nam in AugustTyphoon Maemi - Peak wind speeds of280 km/hr. More than 130 deaths in theRepublic of Korea in September andUS$ 4.1 billion estimated damage

SahelAbove normalrainfall inSeptember.Above normal crops for 2003/04

Tropical Cyclone Tropical Cyclone Ami Crossed Fiji Januarywith wind speeds of185 km/hr

Tropical Cyclone DelfinaDelfina - December/January. 100 km/hr at its peak.Heavy rainfall in Mozambique

Tropical Cyclone Tropical Cyclone Erica Peak winds of 185 km/hr as it crossed New Caledonia in March

Tropical Cyclone Manou - May. 140 km/hrwinds at landfall in Madagascar.265 fatalities

Second lowest September sea-ice extent on record for Arctic. Lowest was in 2002

Warmest summer on recordfor France, Switzerland,Spain and Germany

India and much of AsiaCold January. More than2 500 fatalities

Wildfires and drought insouth-east Australia inJanuary and February

Northern hemisphere snow cover extent. Second greatestsnow cover extent on record for northern hemisphere winter

ArgentinaSanta Fe - Flooding in April/May. Several days of heavy rain. Salado River rose 508 mm in 12 hours

Hurricane FabianHit Bermuda atcategory 3.Caused US$ 300 milliondamage

Hurricane Marty12 deaths. Approximately4 000 homes damaged insouthern Baja California

East Pacific hurricane season16 named storms. 7 hurricanes. First season since 1977 with nomajor hurricanes

ColombiaFlooding inDecember

North America10th lowest winter snowcover extent onrecord for 2002/2003

India, PakistanDeadly heat wave in May/June.Temperatures reached above50°C. 1 500 deaths.

Indian monsoon rainfall102% of normal overall.Distributed evenly

North-east U.S. - February snowstorm. Numerous new24-hour snowfall and storm total records

May tornado outbreak - most tornadoes (412) of any 10-day period. 42 deaths.

MexicoImproving drought conditions in northern Mexico

RRRTemperatures in

n Russia°C

PeruTemperatures reach as low as -20°C in southern Peru in July

Typhoon Krovanh

Typhoon Maemi

Typhoon Imbudo

Typhoon SoudelorTyphoon Etau

France, Portugaland MediterraneanSevere wildfireactivity in Julyand August

Siberia and KazakhstanLarge-scale forest fires inspring and summer

12 typhoons affected China in 2003

AustraliaHeat wave in September. New all-time Australian September record for daily maximum temperature of 43.1°C at West Roebuck

BrazilTorrential rain and flooding in Rio de Janeiro in January

Tropical Storm Linfa

South Africa, BotswanaContinued drought

Antarctic ozone hole largest on record, 28 million km2 in late September

SiberiaAnnual anomalies of +2°C

Southern Ethiopia,Somalia, KenyaWettest conditions in70 years in some areas

Figure 1.1Significant climaticanomalies and events in2003. Average globaltemperature was the thirdwarmest on record.There has been a rise inglobal temperature of0.68°C since 1900.(Sources: NationalClimatic Data Center,NOAA, and WMO)

(+0.48°C), and at the end of arguably thestrongest El Niño of the twentieth century,1998 (+0.55°C).

A detailed description of the weatherand climate of 2003 is given in thefollowing chapters. A summary of thesechapters is as follows:

1.2 Chapter 2: Global climate

Temperature• The global mean surface temperature

anomaly (+0.46°C) from the 1961-1990base period was the third highestobserved during the period ofworldwide instrumental records(beginning in approximately 1880).This resulted in an estimated globalmean temperature of 14.52°C.

• The 10 highest mean annual globaltemperature values have occurredsince 1990, including every year since1997. Since the start of the twentiethcentury, the global averagetemperature has risen between 0.6°Cand 0.7°C. Since 1976, this trend hasbeen closer to 0.18°C per decade.

• Analyses of proxy data for thenorthern hemisphere indicate that thelate twentieth century warmth isunprecedented in the past millennium.

• The global annual average temperatureof both the lower and middle

troposphere was the third warmestsince satellite records began in 1979.Mid-tropospheric temperatures haveincreased at a rate of between +0.04°Cper decade and +0.24°C per decade.

• The global lower-stratospherictemperature was below that of thebase period (1984–1990) for theeleventh consecutive year.

Precipitation

• The worldwide land surfaceprecipitation anomaly was 0.9 per centbelow the 1961–1990 average.

Snow cover

• Annual mean snow cover extent overnorthern hemisphere land areas was25.8 million km2, 0.2 million km2

above the 1967-2003 average (years1968, 1969 and 1971 excluded due tomissing data).

• During the boreal summer of 2003,northern hemispheric snow coverextent was at a record low in July.

1.3 Chapter 3: Trends intrace gases

• The increase in the global climateforcing by greenhouse gases from 2002to 2003 was 0.04 W m-2. Ninety percent of this increase was caused byCO2.

9

1. Executive Summary

Figure 1.2Four tropical cyclones

lined up across thesouthern Indian Ocean

east of Madagascar (left)on 12 February 2003.Cyclones pictured are

Gerry, Hape, 18S, andFiona. (Image courtesy

of Jeff Schmaltz, MODISRapid Response Team,

NASA GSFC)

• CO2 rose approximately 2.5 parts permillion (ppm) to reach a preliminaryvalue of 375.6 ppm at the Mauna LoaObservatory in Hawaii. This large risewas the first time two consecutiveyears had risen by over 2 ppm fromthe previous year. In the southernhemisphere, the Cape GrimObservatory in Tasmania recorded anannual average CO2 value of 373 ppm,a rise of approximately 2 ppm from2002.

• Ozone-depleting gases in thetroposphere continued to declinethrough 2003 in response tointernational measures agreed to by the1987 Montreal Protocol on Substancesthat Deplete the Ozone Layer.

1.4 Chapter 4: The tropics• A moderate warm El Niño/Southern

Oscillation (ENSO) event dissipated

during the boreal spring of 2003 andtransitioned to near-neutral conditionsfor the remainder of the year.

• Atlantic tropical storm activity was aboveaverage in 2003, but below average inthe eastern North Pacific and nearaverage in the western North Pacific.

• Indian Ocean storm activity (Figure1.2) was long lived; stretching overeight months, while near Australiaactivity was slightly below average. Inthe South-West Pacific, an aboveaverage number of storms reachedhurricane intensity during 2003.

1.5 Chapter 5: Polar climate• The Antarctic ozone hole matched the

record extent set in 2002: 28 millionkm2. Ozone-depleting substances inthe stratosphere are thought to havereached their peak. In the loweratmosphere, concentrations of

10

Figure 1.3Smoke from devestatingwildfires raging acrosssouthern Californiaswept out over thePacific Ocean on25 October 2003.Over 300 000 hectareswere burnt and at least20 people killed. Thefires were descibed asthe costliest inCalifornia’s history.(Image courtesy JacquesDescloitres, MODISRapid Response Team atNASA GSFC)

chlorofluorocarbons (CFCs) continuedto decline.

• Arctic sea-ice reached a record lowextent in the boreal autumn.

• Arctic surface air temperatures werewarmer than average, especially inBaffin Bay and the east Siberian Sea.

1.6 Chapter 6: Regional climate

• Africa: Sahel rainfall totals were abovenormal during 2003, yielding recordcrop harvests. Drought continued inparts of the Greater Horn and southernAfrica. In Tanzania, rainfall was 300-500 mm below average, leading tofood shortages for 2 million people.Parts of eastern South Africa,Botswana, Zimbabwe and southernMozambique received only fifth totwentieth percentile rainfall.

• Asia: Monsoonal rains were nearnormal for the Indian subcontinent in2003. Near-normal rainfall helped easethe long-running drought in south-westAsia. In China, flooding in the first halfof the year affected approximately 100million people. Russia was warmerthan average for the year, but June wasanomalously cold in European Russia.

• Australasia: Severe drought througheastern Australia, coupled with extreme

heat during the austral summer, causeddevastating impacts in early 2003. Itwas the warmest June on record forNew Zealand, and numerous annualsunshine records were exceeded acrossthe South Island.

• Europe: A record-breaking summerheat wave over much of westernEurope led to thousands of heat-relatedcasualties. In Britain, temperaturesabove 100°F (37.8°C) were recordedfor the first time in history.

• North America: Drought continuedacross western North America, withimpacts in the western United States(Figure 1.3), Canadian prairies andnorth-eastern Mexico. Anomalously wetconditions impacted the eastern UnitedStates, with record annual precipitationin three mid-Atlantic States.

• South and Central America: Dryconditions since 2001 persisted inCentral America and the Caribbean,with lowest totals in El Salvador,southern Honduras and north-westernNicaragua. Likewise, very dryconditions were experienced acrossparts of both Brazil and Colombia. Incontrast, above average rainfall wasreported in Peru and Ecuador, thelatter resulting in outbreaks of denguefever and leptospirosis. Extreme coldacross the southern reaches of Peruduring July led to over 200 fatalities.

11

1. Executive Summary

12

2.1 Global surfacetemperatures

According to surface temperature recordsmaintained independently by institutions inthe United Kingdom and in the UnitedStates, the global mean surface temperaturein 2003 was within the highest three annualvalues observed during the period ofregular instrumental records (dating to thenineteenth century), but below the 1998record high value. Analyses of proxy datafor the northern hemisphere indicate thatsuch high anomaly values areunprecedented in the past millennium(Jones, et al., 1998; Mann, Bradley andHughes, 1999).

Surface temperatures in 2003 ranked asthird highest, calculated according to theUnited Kingdom record (Jones, et al., 2001;Jones and Moberg (2003)), as shown inFigure 2.1, or tied for second highestaccording to the United States(NOAA/NCDC) record (Quayle, et al.,1999). In the United Kingdom record,global surface temperatures in 2003averaged 0.46°C above the 1961–1990 meanvalue, only slightly (0.02°C) below the 2002value. In the United States, surfacetemperature archive, the global 2003departure from the same base period wasalso calculated at +0.46°C, equivalent to the2002 United States value. These anomalyvalues suggest that the global meantemperature for 2003 was approximately14.52°C (Jones, et al., 1999).

According to both surface temperaturearchives, the 10 highest mean annual globaltemperature rankings have occurred since1990. Although by no means monotonic,the rise in global mean surface temperaturesince 1900 exceeds 0.6°C per century. Since1976, the linear trend is approximately0.18°C per decade.

Averaged over the entire hemispheres,2003 surface temperatures ranked as secondhighest in the northern hemisphere and as

third highest in the southern hemisphere,according to both surface temperaturearchives. While the top 10 annualtemperature rankings in the northernhemisphere are comprised exclusively ofyears since 1990, the southern hemispheretop rankings include 1987 and 1988.

A calendar-monthly mean record-highvalue was observed in September 2003,when the global mean surface temperaturewas calculated to be about 0.05°C higher

Chapter 2

Global climate

(c) Tropics (30°N-30°S)

Temperature anomalySmoothed with abinomial filter

(a) Global

Diff

eren

ce fr

om 1

961–

1990

(°C

)

(b) Northern hemisphere (north of 30°N)

(d) Southern hemisphere (south of 30°S)

1880 1900 1920 1940 1960 1980 2000

-0.4

-0.6

-0.8

-0.2

0

0.2

0.4

0.6

0.8

-0.4

-0.6

-0.8

-0.2

0

0.2

0.4

0.6

0.8

-0.4

-0.6

-0.8

-0.2

0

0.2

0.4

0.6

0.8

-0.4

-0.6

-0.8

-0.2

0

0.2

0.4

0.6

0.8

Figure 2.1Combined annual land(near surface) and sea-surface temperatureanomalies from 1861-2003 (departures indegrees Celsius from theaverage in the 1961-1990 base period) for(a) the globe; (b) thenorthern hemispherenorth of 30°N; (c) theTropics (30°N-30°S);and (d) the southernhemisphere south of30°S. The solid redcurves have hadsubdecadal timescalevariations smoothed withbinomial filter. Anomalies(in degrees Celsius) for2003 are: +0.46 (a);+0.71(b); +0.45 (c);and +0.15 (d).(Source: IPCC, 2001;Hadley Centre, The MetOffice; and ClimateResearch Unit, Universityof East Anglia, UK)

than the former September record, set in1997 (Figure 2.2).

Worldwide, sea surface temperatures(SSTs) in 2003 ranked as the second higheston record. Large regions of positive SSTanomalies, some of which exceeded the90th or 98th percentiles of the mean annualtemperature distribution (Horton, Follandand Parker, 2001) can be seen in Figure 2.3.

The positive temperature anomaliesacross parts of the North Atlantic extendedinto western Europe and North Africa (seesubchapters 6.4 and 6.5.1, respectively),where extreme summer temperatures wereobserved during June–July–August. Largepositive temperature anomalies were alsoevident across much of the northernhemisphere high latitudes, western NorthAmerica, and much of eastern Australia (see

subchapters 5.2, 6.1.2 and 6.7.1,respectively). Across parts of these areas,the 2003 mean surface temperature alsoranked in the upper 10th percentile of theregional temperature distribution.Conversely, negative temperaturedepartures were observed across easternNorth America (see subchapter 6.1.2).

2.2 Upper-air temperatureBulk atmospheric temperatures for threeatmospheric layers, the lower troposphere(LT), mid-troposphere (MT) and lowerstratosphere (LS), were derived from themicrowave emissions of atmosphericoxygen. The University of Alabama inHuntsville (UAH) generated these productsfrom measurements made by microwavesounding units which are flown as part ofthe instrument suite on NOAA polar-orbiting satellites (Christy, et al., 2003).Biases and errors due to spacecraft drift(both horizontal and vertical) and on-orbitcalibration changes have been identifiedand removed.

Two new reconstructions of microwavesounding unit temperatures by Mears,Schabel and Wentz (2003) of remotesensing systems (RSS), and by Vinnikov andGrody (2003), for the MT (as well as the LSby RSS), have been completed in whichdiffering methods for accounting foreast–west drift and instrument calibrationhave been applied, while Fu, et al. (2004)have corrected the microwave soundingunit record for tropospheric contaminationby the stratospheric cooling trend.

2.2.1 Lower troposphere

The annual global average temperature ofthe LT (surface to about 8 km) developed byUAH was +0.20°C above the 1979–1998mean. Seasonal anomalies are shown inFigure 2.4. Relative to the past 25 years,2003 was the third warmest on recordbehind 1998 (+0.47°C) and 2002 (+0.24°C).The linear trend since November 1978 is+0.08°C per decade. Christy, et al. (2003)demonstrated that the 95 per centconfidence level precision for UAH annualglobal LT anomalies was ±0.14°C, and theglobal trend value was ±0.05°C per decade.Global averages from the National Centersfor Environmental Prediction(NCEP)/National Center for AtmosphericResearch (NCAR) reanalyses and the Hadley

13

2. Global climate

Figure 2.2Serial monthly surfacetemperature anomaliesfrom 1961 to 2003,

based on Quayle, et al.(1999). Anomalies were

calculated with respectto the 1961-1990 base

period.

Figure 2.3Geographic distribution

of temperatureanomalies (in °C) in2003: (a) departure

from the 1961-1990average (the temperature

value of each5° lat × 5° long pixel

was derived from atleast one month’s data ineach of four three-monthseasons); (b) same as in

(a), but expressed aspercentiles of modifiedtwo-parameter gamma

distributions fitted toannual data for

1961-1990, andcalculated according to

Horton, Folland andParker (2001).

Centre, both of which are based onradiosonde measurements, also reportedsimilar findings. Regionally, virtually all areaspoleward of 45°N were above the 20-yearmean, while elsewhere scattered areas wereslightly warmer or cooler than average.Globally, the LT seasonal anomalies in 2003(beginning with December–January–February(DJF)) were as follows: +0.28°C, +0.16°C,+0.10°C, and +0.21°C.

2.2.2 Mid-troposphere

Three realizations of MT globaltemperatures from microwave emissionsgive differing results for the past 25 years.UAH and RSS show that 2003 was the thirdwarmest year of the 25-year period ofrecord (Figure 2.5). However, the globaltemperature trend over the 25-year periodis quite different. Linear trend values are+0.04, +0.12, and +0.24°C per decade forUAH, RSS, and Vinnikov and Grody (2003),respectively. The quantity MT is somewhatdifficult to interpret because it receivesemissions from the troposphere as well asthe stratosphere (about 15 per cent), so thatvariations in MT are a combination of oftencounter balancing anomalies above andbelow the tropopause.

2.2.3 Lower stratosphere

The global LS temperature was againsignificantly below the base period(1984–1990), with anomalies of –0.41°C(UAH) and –0.35°C (RSS). These were thecoldest since 2000, making 2003 the 11thconsecutive year in which the annualaverage was below that of the base period(Figure 2.6). Although the bulk of the layermonitored by these microwave emissions isin the lower stratosphere, a significantportion (~20 per cent) of the tropical signalarises from emissions below the tropopauseso that some convolution of upper-tropospheric and lower-stratospheric signalsoccurs there (Fu, et al., 2004). Despite thisfact, the tropical portion of the globeexperienced negative temperatureanomalies throughout 2003, consistent withthe easterly phase of the quasi-biennialoscillation (QBO). The singular warmanomaly in September–October–November(SON) in 2002 over Antarctica (+10°C) didnot reappear in 2003, but the seasonalanomaly was positive (+1.2°C) for the year.

2.3 Global precipitationFor the third consecutive year, globalannual precipitation averaged over landareas was below normal. Figure 2.7 showsthe annually averaged precipitation fromthe Global Historical Climatology Network(GHCN), with anomalies determined withrespect to the 1961–1990 base period(Peterson and Vose, 1997; Vose, et al.,1992). The global precipitation anomalywas negative (dry) in 2003, reachingapproximately -0.90 per cent (-10 mm) fromthe mean. This was a larger departure fromnormal than in 2002, when the observedanomaly was slightly negative, but veryclose to the 1961–1990 average. The stringof negative global precipitation anomaliesbegan in 2001. Prior to this three-yearperiod of negative anomalies and below-normal precipitation, land areas had beendominated by a period of positiveanomalies and wetter-than-averageconditions as measured by the GHCN.Between 1996 and 2000, the globalprecipitation anomaly had been positive(wet) for four of five years.

The recent drier-than-normal globallyaveraged precipitation anomalies may beassociated with changes in ENSO in thetropical Pacific Ocean. It has been shown

14

Figure 2.4Global seasonal lowertropospheric temperatureanomalies (in °C) for85°S-85°N. Anomalieswere determined withrespect to the 1979-1998 base period. Datafrom the University ofAlabama, Huntsville.

Figure 2.5Annual anomalies (in °C)of global mid-tropospheric temperature.Anomalies weredetermined with respectto the 1979-1998 baseperiod.

verification of satellite-based retrievalmethods. A method that has been widelyused is the climate anomaly monitoringsystem–outgoing long-wave radiationprecipitation index (CAMS–OPI), whichuses 1979–1995 as its base period. TheCAMS–OPI precipitation anomalies for 2003are shown in Figure 2.9. Over land areas,the positive and negative anomalies weregenerally similar to that measured byGHCN. Qualitative comparison of GHCNwith CAMS–OPI shows general agreementbetween the two products over most areasin the tropics and mid-latitudes wherecomparable data exist.

Significant positive anomalies (wetconditions) over the oceans as measured byCAMS–OPI in 2003 included a large regionof the tropical western Pacific Ocean, aswell as the western half of the IndianOcean. Negative anomalies (dry conditions)in 2003 included the western North AtlanticOcean and Gulf Stream region, the near-equatorial Pacific basin, the Coral Sea andconvection associated with the South PacificConvergence Zone (SPCZ), and portions ofthe eastern Indian Ocean.

2.4 Northern hemispheresnow cover

Annual snow cover extent over northernhemisphere land areas averaged 25.8 millionkm2 in 2003. This is 0.2 million km2 abovethe three-decade-long average and ranked

16

180° 120°W 60°W 0° 60°E 120°E 180°

90°N

60°N

30°N

0°

30°S

60°S

90°S

–400 –300 –200 –50 100 300 400–100 50 200

Figure 2.9Precipitation anomalies(in mm) for January-December 2003 fromthe CAMS-OPI dataset.Anomalies weredetermined with respectto the 1979-1995 baseperiod. Positiveanomalies (wetconditions) are shown ingreen and yellowshading, while negativeanomalies (dryconditions) are shown inbrown shading.

Figure 2.10February 2003 monthlysnow extent anomalieswith respect toclimatology mapscovering the period fromNovember1966 to May1999. Departures showdifferences in theper cent frequency ofcover between 2003and the long-term mean.

17

2. Global climate

2003 as having the thirteenth most extensivesnow cover over the past 34 years. Thisevaluation included the Greenland ice sheet,as well as the North American and Eurasiancontinents. The total land area covered bysnow in 2003 ranged from a maximum of48.5 million km2 in February, to a minimumof 2.4 million km2 in August. Monthly snowextent values were calculated at RutgersUniversity from weekly maps of snow coverproduced by NOAA.

In February 2003, hemispheric snowcover ranked fourth most extensive overthe past 37 years (Figure 2.10). This

included a ranking of the third largestextent over Eurasia. In November,hemispheric extent ranked eighth over thepast 38 years, with the fourth-most-extensive snow cover on record over NorthAmerica. During the summer of 2003,hemispheric extent was at a record low inJuly, and near or within the lowest tercilefrom June through September.

The rankings for each continent werewithin the same terciles for six months ofthe year, but differed by two terciles inJanuary and October. In both of thesecases, Eurasia was in the upper tercile andNorth America in the lower tercile.

Twelve-month running means ofnorthern hemisphere snow cover extentwere above the long-term averagethroughout the year, especially during thefirst half of 2003 (Figure 2.11). This is thefirst such occurrence since the mid-1990s.Eurasian extent was above average for thefirst time in five years and the observedanomalies were the largest since the mid-1980s, while the North American extent didnot exceed the long-term average in 2003.The mid-1990s was the last time thatnorthern hemisphere extent exceeded thelong-term average, and before that it was inthe mid-1980s.

Figure 2.11Anomalies of monthly

snow cover extent overnorthern hemisphere

lands (includingGreenland) between

November 1966 andDecember 2003. Also

shown are 12-monthrunning anomalies of

hemispheric snow coverextent, plotted on the

seventh month of a giveninterval. Anomalies were

calculated from NOAAsnow maps. Mean

hemispheric snow extentis 25.6 million km2 for

the full period of record.Note that monthly meansfor the period of record

were used for ninemissing months between

1968 and 1971 inorder to create a

continuous series ofrunning means (the

missing months all fellbetween June and

October).

CO2 is removed from the atmosphere bythe oceans and the terrestrial biosphere. Itis interesting to notice that the growth ratein the northern and southern hemispherestations (Figure 3.2) are very similar,emphasizing the global impact of CO2

increases.Variations in the observed atmospheric

increase in CO2 are not due to variations infossil fuel emissions. Most attempts toexplain the interannual variability of theatmospheric CO2 increase have focused onshort-term climate fluctuations (e.g., ENSOand post–Mt. Pinatubo eruption cooling),but the mechanisms, especially the role ofthe terrestrial biosphere, are not wellunderstood. For example, it has beenspeculated that the high CO2 growth rate in1998 was related to unusually warmtemperatures. However, analysis of 1999CO2 measurements indicates a return to anaverage or lower-than-average growth ratein 1999, although 1999 was only slightlycooler than 1998. From 2000 through 2002,the CO2 growth rate has been around1.8 ppm per year, slightly above the long-term average of 1.5 ppm per year.

Understanding the relativecontributions of the ocean and biosphere tothe net global carbon sink has importantenvironmental policy implications. High-precision measurements of atmospheric CO2

abundance alone are insufficient tocalculate this partitioning. Researchers usemeasurements of the carbon-isotopiccomposition of CO2 and atmospheric O2 toconstrain understanding of this partitioning.The average CO2 uptake for the 1990s hasbeen estimated using these methods to be1.7±0.5 GtC per year by the oceans and1.4±0.7 GtC per year by the terrestrialbiosphere.

3.2 Methane (CH4)The contribution of CH4 to anthropogenicradiative forcing, including direct andindirect effects, is about 0.7 W m-2,approximately half that of CO2. The indirecteffects include changes in the burden ofCH4, which feedback into atmosphericchemistry affecting the concentrations ofhydroxols (OH) and O3. The increase inCH4 since the pre-industrial era isresponsible for about one-half of theestimated increase in backgroundtropospheric O3 during that time. Changesin OH concentration affect the lifetimes ofother greenhouse gases, such as thereplacement refrigerants (halogenatedfluorocarbons (HFCs) and halogenatedchlorofluorocarbons (HCFCs)).

High-precision measurements ofatmospheric CH4 provide climate modelerswith current and past rates of CH4 increase.They are also useful in constraining the CH4

budget and how it is changing with time. InFigure 3.3a, smoothed, globally-averagedCH4 mole fractions from the NOAA/CMDLair-sampling network are plotted as afunction of time. During nearly 20 years ofmeasurements CH4 has increased, but therate of increase has slowed (Figure 3.3b). Infact, at the Cape Grim station, CH4 hasessentially stopped accumulating in theatmosphere over the past four years. Recentanalyses have indicated that globallyaveraged CH4 remained about constantfrom 1999 to 2002 (Dlugokencky, et al.,2003).

It is reasonable to question whetherthis observed CH4 trend is temporary or anindication that atmospheric CH4 hasreached a steady state. Based on currentknowledge of the global CH4 budget it isimpossible to tell. However, there is good

19

3. Trends in trace gases

Figure 3.2a(a) monthly mean CO2

mole fractions (ppm)from the NOAA MaunaLoa Observatory on the

Big Island of Hawaii.Prior to May 1974,

monthly mean CO2 dataare compliments of C.D.

Keeling, ScrippsInstitution of

Oceanography (bluesymbols), and since May

1974, data arecompliments of

K. Thoning,NOAA/CMDL (red

symbols); (b)monthly mean baseline

CO2 concentrations(ppm) from the Cape

Grim Baseline AirPollution Station,

Tasmania, Australia.(Source: CSIRO

Atmospheric Researchand Australian Bureau ofMeteorology/CGBAPS)

21

3. Trends in trace gases

HCFCs did not increase as rapidly in 2003compared to earlier years. By mid-2003, thechlorine in the three most abundant HCFCsamounted to 205 ppt, or 7.4 per cent of allchlorine carried by long-lived, purelyanthropogenic halocarbons. Total chlorinefrom the HCFCs increased at 8 ppt per yearover this two-year period, which is slightlyslower than the mean rate of increase of 9.3ppt per year observed since 1992. Mixingratios of methyl chloroform continued todecline at an exponential rate; theabundance of this gas has declined nearlysixfold since the early 1990s.

The understanding of atmospherictrends for ozone-depleting gases improvedduring 2003 with the publication of recenttrends for methyl bromide (CH3Br)(Montzka, et al., 2003). Methyl bromide is agas with substantial natural andanthropogenic sources. The global mean of8.1 (±0.1) ppt estimated from measurements

at 10 sites during 2002 was about 1.2 (±0.3)ppt (or about 10 per cent) lower than themean during 1995–1997. The global meanrate of change from mid-1999 to mid-2002was –0.35 (±0.05) ppt per year, which wasa sharp contrast to the continuous increasesreported for atmospheric CH3Br throughoutmost of the twentieth century (WMO, 2003).The observed decrease coincided withdeclines in industrial production, but wassubstantially faster and larger thanexpected.

The sum of chlorine and bromineatoms in long-lived trace gases provides anestimate of the ozone-depleting power ofatmospheric halocarbons after the enhancedefficiency of bromine to destroy ozone,compared to chlorine, has been considered(a factor of 45 is used here; Daniel, et al.,1999). This sum is often expressed asequivalent chlorine (ECl). Until recenttrends of methyl bromide becamepublished, assessments of changes in EClincluded the assumption that mixing ratiosof this gas were constant over time (Figure3.6; blue line). With this assumption, thedecline in ECl was estimated atapproximately 0.5 per cent per year since1995 (Montzka, et al., 1999; Hall, et al.,2002). With measurements of methylbromide now available, however, it wouldappear that ECl has declined 50 per centfaster since 1998 than previously thought(Figure 3.6, red line).

Figure 3.5Globally averaged

surface tropospheric CO(symbols) determined

from the CMDLcooperative air-sampling

network. Red solid linedenotes the de-

seasonalized trend.(Courtesy of P.C.

Novelli, NOAA/CMDL)

Figure 3.6Changes in global

tropospheric mixing ratios(in ppt, or pmol mol–1) ofthe most abundant CFCs

(in purple), HCFCs (inred), chlorinated solvents,and brominated gases (in

green; halons andCH3Br). These global

changes are estimatedfrom weighted means of

measurements from flasksand on-site instruments at

8-10 sites (theNOAA/CMDL

cooperative air samplingnetwork). Appearing in

the bottom panel areglobal tropospheric

trends for ECl, which isthe sum of [chlorine +

(45 × bromine)] from allthe trace gas

measurements displayedin the upper panels (red

line) and from thosesame gases but with the

assumption that methylbromide mixing ratios

have remained constantover time (purple line).

(Updated from Montzka,et al., 1999 and 2003).

(Courtesy ofS.A. Montzka, J.H.

Butler, T. Thompson,D. Mondeel, and

J. Elkins, NOAA/CMDL)

22

4.1 ENSO and the tropicalPacific

4.1.1 Overview

An equatorial Pacific warm episode (ElNiño) occurred from early in the borealsummer 2002 through to early in the springof 2003 (Figure 4.1). This event wasconsistent with past warm episodes in thatit featured anomalously warm watersextending to approximately 100–200 mdepth across most of the central equatorialPacific, in association with a deeper-than-average oceanic thermocline. Increasedthermocline depths are indicated bypositive depth anomalies of the 20°Cisotherm, which approximates the centre ofthe thermocline across the easternequatorial Pacific (Figure 4.1a). Thisanomalous warmth was consistent withreduced oceanic upwelling across thecentral and east-central equatorial Pacific, inresponse to a weakening of the low-leveleasterly winds (Figure 4.1b, orange).

The largest area-averaged SSTdepartures (+1.75°C) associated with thisevent occurred during October 2002–January2003 (Figure 4.1c). During the second half ofOctober 2002, strong equatorial westerlywind anomalies, associated with theMadden–Julian Oscillation (MSO), triggeredan eastward-propagating oceanic Kelvinwave which contributed to an increase inthermocline depths and a sharp increase inSST anomalies in the central and east-centralequatorial Pacific (Halpert and Bell, 2003;see their section 4a).

During this period, enhancedconvection was over the central equatorialPacific (Figure 4.1e, orange) and suppressedconvection across Indonesia and north-eastern Australia. The appearance of above-normal heights at 200-hPa (Figure 4.1d,orange) reflected the increase in deeptropospheric heating over the centralequatorial Pacific in response to theenhanced convection.

4.1.2 Equatorial Pacific Ocean sea-surface and sub-surfacetemperature evolution

During December 2002–February 2003,anomalously warm SSTs (exceeding 29°C)covered the central equatorial Pacific, withthe centre of SST anomalies in the Pacificbasin shifted east of the Date Line. PositiveSST departures of +1° to +2°C extendedfrom the Date Line to 120°W, anddepartures of +0.5° to +1.0°C covered theeastern equatorial Pacific (Figure 4.2a).

During March–April–May (MAM) theeastern equatorial Pacific cold tonguebecame stronger than average (Figure 4.2c),while temperatures west of the Date Linewarmed to more than 0.5°C above average.This evolution marked the end of the ElNiño episode. During June-July-August(JJA) and SON, the largest positive SSTanomalies (Figures 4.2f, h) established westof the Date Line, and near-normal

Chapter 4

The tropics

Figure 4.1Monthly time series of (a)the depth of the 20°Cisotherm (m); (b) 850-hPa zonal wind speed(m s-1); (c) SST (°C); (d)200-hPa height (m); and(e) OLR (W m-2) over thecentral equatorialPacific. Values weredetermined by averagingover the region boundedby 5°N-5°S and 180°-100°W in (a-d), and20°N-20°S and 180°-100°W in (e). The solidcurve in all panels showsthe monthly mean valuesand the dashed curvedshows the climatologicalmean. The anomaliesare shaded, with orangeshading indicatingpositive anomalies andblue indicating negativeanomalies, except for in(e) where the shadingconvention is reversed.The climatology andanomalies werecomputed with respect tothe 1979-1995 baseperiod.

temperatures prevailed across the east-central and eastern equatorial Pacific.

The subsurface thermal structure is afundamental component of ENSO. DuringDJF, anomalously warm waters were evidentbetween 100- and 200-m depths across thecentral and east-central equatorial Pacific(Figure 4.3a). During MAM, this structurechanged markedly as the thermoclinebecame shallower than average across theeast-central and eastern equatorial Pacific,and anomalously cold water developed over

the eastern Pacific between the surface andthe 150-m depth (Figure 4.3b). This cold-episode like structure of the oceanicthermocline did not persist, however, andslightly warmer-than-average waters returnedduring JJA and SON (Figures 4.3c, d).

4.1.3 Atmospheric circulation

In the Tropics, the El Niño–relatedatmospheric features included: (a) lower-level westerly wind anomalies across theeastern equatorial Pacific in association witha reduced strength of the tropical easterlies(Figures 4.4a); (b) an enhanced ascendingmotion and convective activity over thecentral equatorial Pacific; and (c) ananomalous descending motion andsuppressed convective activity in the vicinityof north-eastern Australia (Ropelewski andHalpert, 1987). This reduced strength of theequatorial Walker circulation has been awell-known characteristic of Pacific warmepisodes.

In the meridional direction, thecirculation during DJF 2002/2003 featuredanomalous poleward flow in the uppertroposphere over the central tropical Pacificextending into the core of the East Asian jetstream. This suggests that the Hadleycirculation was enhanced by the warm event(Mo and Kousky, 1993).

Also typical of past warm episodes, theDJF 2002/2003 season featured upper-levelanticyclonic stream function anomalies in the

23

4. The tropics

Figure 4.2Seasonal SST (left) andanomaly (right) for (a),(b) DJF 2002/2003;

(c), (d) MAM 2003; (e),(f) JJA 2003; and (g), (h)

SON 2003. Contourinterval is 1°C, with

0.5°C anomaly contourincluded. Anomalies are

departures from the1971-2000 adjustedoptimal interpolation

climatology (Smith andReynolds, 1998).

Figure 4.3Equatorial depth-

longitude sections ofocean temperature

anomalies for (a) DJF2002/2003; (b) MAM

2003; (c) JJA 2003;and (d) SON 2003.

Shading interval is 1°C,and the dark line is the

20°C isotherm. Datawere derived from an

analysis system thatassimilates oceanic

observations into anoceanic general

circulation model(Behringer, Ji andLeetmaa, 1998).

Anomalies aredepartures from the

1981-2000 base periodmeans.

subtropics of both hemispheres, indicated bypositive (negative) anomalies in the northern(southern) hemisphere, which flanked theregion of enhanced convection over thecentral equatorial Pacific (Figure 4.5b) (Moand Kousky, 1993). These anticyclonicanomalies reflected a strengthening andeastward extension of the subtropical ridgesin both hemispheres, and an overallweakening of the mid-Pacific troughs farthereast. Enhanced westerlies and anomalousupper-level confluent flow across the centralextratropical Pacific were evident along thepoleward flanks of the anomalous subtropicalridges in both hemispheres (Figure 4.5a).These anomalies reflected an eastwardextension of the mid-latitude jet streams inboth hemispheres and an eastward shift inthe areas of strong upper-level diffluence thatdefined the jet exit regions.

These El Niño conditions dissipatedduring MAM 2003 as anomalous cross-equatorial flow at 850-hPa developed acrossthe Pacific (Figure 4.4b), and resulted inenhanced oceanic upwelling and a rapidcooling of ocean temperatures across theeastern Pacific.

4.2 Tropical storms4.2.1 Atlantic hurricane season

The North Atlantic hurricane seasonofficially runs from June through November.An average season produces 10 tropicalstorms, six hurricanes and two majorhurricanes, defined as maximum sustainedwind speeds at or above 100 kts, andmeasured by categories 3–5 on theSaffir–Simpson scale (Simpson, 1974)). In

2003, the Atlantic basin was extremelyactive, with 16 tropical storms, sevenhurricanes and three major hurricanes.

Five of these Atlantic storms madelandfall in the United States, one as atropical depression (Henri), two as tropicalstorms (Bill and Grace), and two ashurricanes (Claudette and Isabel). A sixthsystem, hurricane Erika, made landfall innorth-eastern Mexico, and brought tropicalstorm–force winds and precipitation tosouthern Texas. Mexico also experiencedtropical storm conditions from Claudetteand Larry, and tropical storm Odetteaffected Hispañola. Nova Scotia andBermuda experienced devastating impactsfrom hurricanes Juan and Fabian,respectively.

24

Figure 4.4OLR anomalies (shadedand 850-hPa vectorwind anomalies andisotachs for (a) DJF2002/2003 and (b)MAM 2003. Contourinterval for isotachs is2 m s-1. Shading intervalfor OLR anomalies is 10W m-2. Anomalies aredepartures from the1979-1995 base periodmonthly means.

Figure 4.5DJF 2002/2003: 200-hPa (a) vector wind andisotachs and (b) streamfunction anomalies.Contour interval forisotachs in (a) is4 m s–1. Shadinginterval for streamfunction anomalies is3 × 10 m2 s–1. In thenorthern hemisphere,positive (negative) streamfunction anomaliesindicate an anticyclonic(cyclonic) circulation. Inthe southern hemisphere,negative (positive) valuesindicate an anticyclonic(cyclonic) circulation.Anomalies aredepartures from the1979-1995 base periodmonthly means.

Most of the activity during Atlantichurricane seasons occurs duringAugust–October, primarily in response tosystems developing from African easterlywave disturbances. During the above-normal 2003 season, 10 tropical storms, ofwhich four became hurricanes, developedbetween mid-August and mid-October.Three of these systems became majorhurricanes. Above-normal hurricane seasonsalso feature a high concentration of activityin the main development region (MDR)(Goldenberg and Shapiro, 1996), whichconsists of the tropical Atlantic andCaribbean Sea between 9° and 21.5°N (seemap inset in Figure 4.6). Eight tropicalstorms formed in the MDR during 2003;four of these systems became hurricanes,with three becoming major hurricanes.

Another notable aspect of the seasonwas the formation of five tropical stormsover the Gulf of Mexico, which was equalwith the season high observed in 1957. Onaverage, one to two tropical storms form inthis region during a given season. Also,three tropical storms formed outside thenormal (June–November) hurricane seasonin 2003. Tropical storm Ana formed on 22April, and tropical storms Odette and Peterformed on 4 and 9 December, respectively.This was the first season since 1887 that twotropical storms have formed in December.

Important aspects of the atmosphericcirculation during the peak of the 2003

season (Figure 4.7) can be attributed to theongoing active Atlantic multidecadal signal(Chelliah and Bell, 2004), including: (a) anamplified subtropical ridge; (b) reducedvertical wind shear in the MDR resultingfrom upper-level easterly wind anomalies(green arrows) and lower-level westerlyanomalies (light blue arrows); (c) anexceptionally favourable African easterly jet(dark blue arrow); (d) an active WestAfrican monsoon system; and (e) above-average SSTs in the MDR. During August,the exceptionally conducive nature of thetotal signal was also related to a pre-existing mid-latitude circulation patternknown as the positive phase of the EastAtlantic teleconnection pattern, and duringSeptember–October it was related to ananomalous atmospheric warming across theentire tropical Atlantic in association with abroader warming of the global tropicalatmosphere.

Total seasonal activity may bequantified with the accumulated cycloneenergy (ACE) index (Bell, et al., 2000). Thisindex is a wind energy index, calculated bysumming the squares of the estimated six-hourly maximum sustained wind speed inknots (Vmax2) for all periods while thesystem is either a tropical storm orhurricane (Figure 4.6, black bars). The totalACE index for the 2003 season was174.75 × 104 kt2, or 200 per cent of the1951–2000 median value (87.5 × 104 kt2).Hurricane Isabel, which produced one ofthe largest observed ACE values(63.3 × 104 kt2) of any Atlantic hurricane onrecord, lasted eight days as a majorhurricane and 1.75 days at category 5 status(wind speeds at or above 140 kts).

4.2.2 Pacific tropical storms

WESTERN NORTH PACIFIC TYPHOON

SEASON

Typhoons can develop throughout the yearin the western North Pacific, with peakactivity usually between July and November.In 2003, there were 28 disturbances in thewestern North Pacific basin, includingtyphoons, tropical storms, and tropicaldepressions. This was slightly below the1973–2002 average of 30.3 (median 30.5)(Joint Typhoon Warning Center (JTWC)best-track dataset; JTWC, 2004). One ofthese (Jimena) developed in the easternPacific and entered the western Pacific with

25

4. The tropics

Figure 4.6Seasonal values of the

ACE index for the entireAtlantic basin (black)

and for the MDR (red).The MDR consists of the

tropical Atlantic to21.5°N and the

Caribbean Sea (seeinset). The ACE index for

the MDR is based onsystems that first became

tropical storms in thatregion. NOAA definesneat-normal seasons as

having a total ACE valuein the range of

65 103 × 104 kt2(green lines).

Figure 4.7Schematic representationof conditions during the

peak (Aug-Oct) of theabove-normal 2003

Atlantic hurricaneseason.

tropical depression intensity, while all of theothers formed west of the Date Line (180°).There were 23 tropical cyclones with at leasta tropical storm intensity (referred to as“named tropical cyclone”), two of which(03W and 27W) were not named due totheir minimal intensity—below the1973–2002 average of 26.4 (median 26, 25thpercentile 24; Figure 4.8a).

The number of tropical cyclonesreaching typhoon intensity (17) was closeto the 1973–2002 average of 16.5 (median16). Five typhoons (Kujira, Imbudo,Maemi, Parma and Lupit) reachedsupertyphoon intensity, above the1973–2002 average of 3.6 (median 3, 75thpercentile 5; see Figure 4.8b). On averageover the last 30 years, 12 per cent of alltropical cyclones reached supertyphoonintensity. In 2003, the percentage wasslightly higher (18 per cent).

The ACE index for the whole year wasslightly above normal (not shown), with avalue of 327.6 × 104 kt2, which correspondsto 122 per cent of the median value(268.3 × 104 kt2) of the last 30 years(1973–2002). Overall, tropical storms andtyphoons in the western North Pacificduring 2003 had an average tropicalcyclogenesis position of 12.8°N, 137.5°E,which was west of the 30-year historicalaverage (12.8°N, 143.4°E). The averagetrack position of all named tropicalcyclones in 2003 was 18.6°N, 131.7°E,which was also west of the historicalaverage position (19.0°N, 134.2°E) observedover the last 30 years.

The 2003 typhoon season producedsignificant impacts in many countries of theregion. Among the typhoons that madelandfall, supertyphoon Maemi (5–13September) caused widespread destructionin South Korea. Maemi formed east of thePhilippines and moved north-westwardwhile intensifying, reaching its maximumintensity near Okinawa, Japan, andweakened before making landfall in Koreaand finally modifying into an extratropicalsystem in the Sea of Japan. High winds andrainfall associated with Maemi wereresponsible for more than 100 deaths anddestroyed more than 1.4 million houses,damaged roads and bridges, and sank atleast 82 vessels.

EASTERN NORTH PACIFIC HURRICANE

SEASON

The hurricane season in the eastern NorthPacific Ocean typically begins in mid-May

and extends through to the end ofNovember. Reliable records for the easternNorth Pacific (ENP) basin date toapproximately 1951.

In 2003, the hurricane season in theENP basin was below normal in severalnotable categories. The 2003 seasonincluded a total of 16 named storms, sevenhurricanes and no major hurricanes. Thetotal number of named storms in 2003equalled the climatological average of 16for the ENP basin, while the formation ofseven hurricanes was below the long-termaverage of nine. No major hurricanesformed during the entire season, while onaverage four typically develop. This was thefirst year since 1977 where no category 3–5hurricanes developed during the entiretropical cyclone season in the ENP basin.

The 2003 season was also notable dueto the record late development of theseason’s first hurricane. On 24 August,hurricane Ignacio intensified from a tropicalstorm to a category 2 hurricane. Theaverage date for the development of thefirst hurricane of the season in the ENP is24 June.

On average, 1.34 tropical storms and1.3 hurricanes make landfall during the ENPhurricane season (15 May–30 November)along the Pacific coast of Mexico, withrespect to the 1950–2000 mean (Jauregui,2003). Overall, four of the 16 named stormsmade landfall during the 2003 ENPhurricane season (Figure 4.9). Two of these,Ignacio and Marty, intensified into

26

Figure 4.8Number of (a) namedtropical cyclones and (b)supertyphoons per yearin the western NorthPacific basin during theperiod 1950-2003. Thesolid green line indicatesthe median for 1973-2002, and the dashedgreen lines show the25th and 75thpercentiles. (Datacourtesy of JTWC)

hurricanes and crossed the Baja Peninsulain Mexico, while tropical storms Carlos andOlaf struck the Pacific coast of Mexico. Onelandfalling tropical storm, Nora, was notcounted in the total for the season becauseit made landfall along the central Mexicancoast only after it had weakened to adepression in early October. Therefore,despite the below-average number ofhurricanes in 2003, the number oflandfalling tropical storms and hurricaneswas above average for the season.

The lack of any major hurricanes, thedevelopment of only one category 2hurricane, as well as the record lateformation of the first hurricane of theseason made 2003 a below-normal year inthe ENP basin. This was strongly reflectedin the ACE index (Bell, et al., 2000) for the2003 ENP hurricane season. The ACE indexfor the 2003 season was 48.67 × 104 kt2

(G.D. Bell, 2004, personal communication),which was approximately 37.5 per cent ofthe mean value of 129.7 × 104 kt2 andapproximately 43 per cent of the medianvalue of 113 × 104 kt2 determined from the1971–2000 base period climatology. Severalenvironmental factors influenced theobserved anomalous conditions, and theseincluded: ENP basin SSTs, higher totalvertical wind shear over the ENP, the phaseof ENSO, and the phase of the QBO in thetropical lower stratosphere.

AUSTRALIA AND THE SOUTH-WEST

PACIFIC

During 2003, there were 10 tropicalcyclones in the Australian region (south ofthe Equator between longitudes 90°E-160°W), which is slightly below the 22-yearannual average of 11.5. Similarly, ninetropical cyclones impacted upon the South-West Pacific. One cyclone (Erica) affectedboth regions.

In the Australian region, almost all thecyclones began within monsoon or near-equatorial troughs, except for tropical

cyclone Erica, which began as a low overthe Australian continent that movedeastwards into the Coral Sea where itdeveloped into a cyclone. In fact Erica wasthe only cyclone to affect the Coral Seaduring the year, which is substantiallybelow the average of four. Four cyclonesdeveloped in the Arafura Sea/Gulf ofCarpentaria area, while five cyclonesaffected the Indian Ocean. Five systems,Fiona, Erica, Inigo, Jana and Debbie,attained severe tropical cyclone status.

Six of the 10 cyclones made landfallover the Australian continent. Of these,three made landfall along the north-westcoast. Tropical cyclone Inigo also affectedIndonesia during its development phase,producing heavy rainfall that causedlandslides and flooding and killed at least50 people. While a number of cyclonesproduced heavy rainfall upon landfallcausing significant flooding, with manyroads being cut by swollen rivers, damagewas generally light with only one fatality,the result of floodwaters in far north-westAustralia from tropical cyclone Graham.

In the South-West Pacific, eight tropicalcyclones occurred during the calendar year,all of which occurred during the 2002-2003tropical cyclone season. Interestingly, thetwo tropical cyclone seasons that overlapthe 2003 calendar year were vastly different.The 2002-2003 tropical cyclone season wasabove the long-term (1970/1971-2000/2001)mean of nine, with 11 tropical cyclonesnamed during the season, while the 2003-2004 season only had four systems, equalto the record low year of 1994-1995. The2003 calendar year saw six tropical cyclonesreach hurricane intensity, which exceedsthe climatological base period seasonalaverage of between four and five.

During early 2003 the impact of activeMadden-Julian Oscillation (MJO) pulsesacross the South-West Pacific generallybrought strong convective activity andcontributed towards tropical cyclonedevelopment. SST anomalies were nearnormal in the Coral Sea region throughoutthe season.

Of the nine tropical cyclones to affectthe South-West Pacific during the calendaryear, two caused severe damage and loss oflife. Tropical cyclone Ami (12-15 January)brought hurricane force winds to Fiji’sNorthern and Central Divisions, anddamaging gale force winds affected Tongaand Tuvalu. Damage in Fiji was extensiveand severe due to high winds, heavy seas

27

4. The tropics

Figure 4.9The observed tracks of

all named tropicalstorms and hurricanes inthe eastern North PacificOcean during the 2003

season. (Data courtesyof the University of

Hawaï)

and torrential rainfall that led to severeflooding in the northern town of Labasa.The total confirmed number of fatalitieswas 17. Damage was severe in the social,economic, infrastructure and utilitiessectors, particularly in the Macuata,Cakaudrove and Lau Provinces. Theestimated cost of damage inflicted by Amiwas FJD 104 million. Tropical cyclone Erica(3-16 March), which started life in Australia,struck New Caledonia as it carved eastwardover the Coral Sea. Erica reached its peakintensity of 115 knots and left up to 1 000homeless in New Caledonia as well asseriously damaging public buildings. Twofatalities were reported with nine peopleseriously injured. Erica was arguably theworst tropical cyclone to affect NewCaledonia in 50 years, causing damageestimated up to CFPF 450 million.

An interesting aspect of the South-WestPacific year was the formation of twotropical cyclones (Gina and Epi) after theofficial end of the tropical cyclone season.Both were relatively weak and short lived,however the occurrence of not one, buttwo tropical cyclones in the South-WestPacific during June was highly unusual.

4.2.3 South-west Indian Oceantropical storms

As for Australia and the South-West Pacific,the south-west Indian Ocean tropicalcyclones for 2003 include tropical cyclonesfrom the end of the 2002-2003 season andthe start of the 2003/2004 seasons.

The 2002-2003 cyclone season wasonce again long and active in the south-west Indian Ocean Basin. Thirteendisturbances were classified as havingreached the stage of tropical storm (asopposed to 11 the previous season), onlyone storm less than the maximum of 14recorded since the beginning of the satelliteera (1967-1968). Eight of these stormsreached hurricane force. The seasonalaverage in the basin is nine tropical storms,of which slightly less than half reach thecyclone stage.

The high number of tropicaldisturbances in the region did not stemfrom an unusually high incidence ofcyclogenesis. Sixteen systems resulted inadvisories being issued, which is the samenumber as during the 1997-1998 season oflow activity. It is interesting to note that thecyclogenesis rate has been very stable

during recent seasons, staying within the14-16 range for five years, with the soleexception for the 2000-2001 season. Therate of development into maturedisturbances (13 tropical storms for 16embryonic disturbances) was particularlyhigh, generally indicating very favourableenvironmental conditions.

The number of days of cyclone activitymay be a more representative indicator ofthe strength of the season. With a total of24 cyclone days, the season was 20 percent above the norm (the mean is 20 days)and lags far behind the 35 days of 2001-2002. Just one storm managed to sustainmore than three days at hurricane force(Kalunde), as opposed to four which didso the previous season. With 13 tropicalstorms or cyclones, a total of 68 days wererecorded as having a system of at leastmoderate tropical storm strength. This totalis considerably higher than normal (meanof 53 days, median of 48).

The 2002-2003 tropical cyclone seasonin the south-west Indian Ocean stretchedout over more than eight months. Since thebeginning of the satellite era (1967-1968),only four seasons have started earlier andfour have finished later.

Very little populated land was able toescape the influence of a tropical cyclonein the calendar year 2003. Southern Africaand especially the Mozambique coastlinefelt the impact of tropical storm Delfina inthe first days of the year, then tropicalcyclone Japhet in late February/earlyMarch. Rains and floods associated withthese two events claimed dozens of lives inMozambique and Malawi.

The Mascarenes saw severaldisturbances move in their vicinity.Mauritius was only moderately affected bytropical cyclone Gerry in mid-February,while the island of Rodrigues was not solucky, suffering the fierce winds of tropicalcyclone Kalunde on Independence Day inMauritius, March 12. This category 5cyclone was the strongest of the season,and hence despite being in a weakeningphase when it struck Rodrigues, gusts ofmore than 200 km h-1 were recorded.Torrential rain, with up to 300 mm beingrecorded in eight hours, caused significantdamage on the small Mauritian island.

At the end of the season, Madagascar,which had only experienced the verymoderate influence of the tropical stormFari in late January, paid a heavy toll whenthe last system of the season, tropical

28

cyclone Manou (Figure 4.10) hit the eastcoast of the ‘big island’ on 8 May, claimingnumerous victims.

Since 1967 only six tropical cycloneshave been recorded over the basin duringMay, of which Manou was the mostsoutherly (19ºS). Furthermore, Manou alsomarked the second year in a row in whicha tropical cyclone had hit populated landduring May; the first time this has everoccurred.

The 2003-2004 tropical cyclone seasonstarted with another very early system, withthe first tropical cyclone, Abaimba, formingin late September east of the Seychelles.

The final four cyclones of the calendaryear resulted in no significant damage orloss of life. For the calendar year 2003, atotal of 15 storms occurred in the south-west Indian Ocean, with eight havingreached hurricane force. This is equal tothe record number of storms observed in astandard south-west Indian Ocean tropicalcyclone season (1 July–30 June).

29

4. The tropics

Figure 4.10On 8 May 2003,

tropical cyclone Manoumade landfall on

Madagascar. Imagefrom the moderateresolution imagingspectroradiometer

(MODIS) onboard theAqua satellite. (Image

courtesy of JacquesDescloitres, MOSDIS

Rapid Response Team atNASA GFSC)

5.1 Antarctic5.1.1 Surface climate

Reliable surface measurements are diffi-cult to obtain for Antarctica given theharsh conditions and transience of scien-tific personnel, however from the limiteddata available the last decade has beencolder than average. This is in contrast tothe general increasing temperature trendbased on data during the late 1950s andthe early 1990s. The June 1991 eruptionof Mt. Pinatubo may have had a markedinfluence on all aspect of Antarctic tem-peratures for several subsequent years(Jacka and Budd, 1998). Overall, 2003was cooler than 2002, but close to the 10-year mean.

Antarctic sea-ice covers approximate-ly 17 to 20 million km2 of the SouthernOcean at its maximum in late winter/earlyspring. By late summer/early autumn,only about 3 to 4 million km2 remains,compared with approximately 7 to 9 mil-lion km2 in the Arctic. As shown in Figure5.1a, in 2002, sea-ice extent in theAntarctic was below average for much ofthe year. By year’s end, sea-ice extent hadbegun to increase in comparison to the30-year average, and for the first sixmonths of 2003, sea-ice extent was aboveaverage. A brief decrease relative to themean was observed in early to mid-winter,and then a recovery to above-average sea-iceextent occurred during the remainder of2003. Sea-ice surrounding the Antarctic conti-nent has been increasing from the late 1970sto the mid-1990s (Figure 5.1b), but this

30

Chapter 5

Polar climate

Figure 5.1Monthly sea-ice extentanomalies (in 106 km2)for (a) the northernhemisphere for 2001-2003; (b) the northernhemisphere from 1973to 2003; (c) the southernhemisphere for 2001-2003 and (d) thesouthern hemisphere from1973 to 2003.Anomalies are departuresfrom the 1973-1996base period.

time series belies considerable seasonaland spatial variability. The early part(1973–1978) of this sea-ice record isdetermined by using different instrumentsand may reflect data calibration problems(Folland and Karl, 2001a).

5.1.2 Antarctic stratospheric ozone

The life cycle of the 2003 ozone hole wassimilar to that of the record sized holeobserved in 2000 (Figure 5.2). By August2003 conditions in the stratosphere wereprimed for another large ozone hole.Temperatures in the lower stratospherewithin the polar vortex were at the lowend of the range measured during theprevious 20 years, with temperatures aslow as –100°C being reached early in themonth. By mid-September, this outlookproved correct, with the ozone holereaching 28 million km2. A brief decreasein size occurred, but then the holeexpanded once more and again reached28 million km2 in late September, havingvaried between 25 and 28 million km2

throughout the month. This peak area iscomparable to the record ozone holeobserved in 2000 (28.5 million km2) andfar exceeds the 1991-2002 mean peakvalue of over 21 million km2. Of someconcern was the area of ozone values,which were 50 per cent below their pre-ozone-hole norms (i.e., ozone values dur-ing the period 1964-1976). This 50 percent area reached 15 million km2 for thefirst time in recorded history, eventually

peaking on September 26 at 18 millionkm2. The 50 per cent area had onlybreached 10 million km2 four times previ-ously.

By the second week of October, theozone hole had reduced to 18 millionkm2, with a rapid reduction in size untilits disappearance in mid-November. Incontrast, in 2000 the hole did not dissi-pate until mid-December. Overall, thepeak ozone hole size was not only com-parable to the 2000 record, but was alsosome 9 million km2 larger than the ozonehole of the previous year, when theozone hole effectively split in two in lateSeptember.

In further contrast to 2002, the ozonehole was far deeper in 2003, reaching aminimum total column ozone of around100 Dobson Units (DU) in lateSeptember/early October, considerablyless than the 2002 value of 135 DU, butthankfully more than the record satelliteobservation of 88 DU in 1994. The select-ed South Pole profiles from 6 August and26 September (Figure 5.3b, c) show thatozone in the 14–21 km layer was nearlycompletely destroyed by the time theminimum total ozone column ozone atthe South Pole of 106 DU was observedin mid-September. The 60 per cent dropin total column ozone was equal to the10-year average loss of 60 ±6 per cent.

The return to near record ozone val-ues in 2003 highlights the commentsmade the previous year (WMO, 2002) thatthe relatively weak nature of the 2002

31

5. Polar climate

40

35

30

25

20

15

10

5

0

Are

a of

ozo

ne h

ole

(mill

ion

km2 )

DecemberNovemberOctoberSeptemberAugust

2000 2001 2002 2003

Figure 5.2Daily size of the Atlantic

ozone hole from1 August to

30 November for theperiod 2000-2003.

(Source: Ozone dataanalysis are prepared in

collaboration with theWMO World Ozone

and Ultraviolet DataCentre in Toronto,

Canada through thecooperation and support

of the MeteorologicalService of Canada)

ozone hole was a result of anomalous cir-culation and meteorological conditionsassociated with the polar vortex and notan indication of a return to a normalchemical composition of the stratosphere.

5.2 The ArcticTemperatures over the Arctic basin for2003 were warmer than average (Figure5.4a), with the boreal autumn and winterexhibiting the largest anomalies over thewestern half of the Arctic Ocean andmoderate cool anomalies over the eastern

part of the basin. This broadly coincidedwith the pattern of positive and negativeheight anomalies at the 500-hPa level inthe region (Figure 5.4b). The borealspring and summer were moderatelywarm over most of the basin.

Temperature trends over the last 50years for the Arctic illustrate significantwarming across the northern continents.However, it is unclear what trend hasoccurred across much of the central ArcticOcean due to lack of data. Much of the50-year trend in the land-based annualtemperatures was derived from northern

32

Figure 5.4(a) Mean annual surfaceair temperatureanomalies (in °C) for2003, with respect tothe 1954-2003 mean;(b) mean annual 500-hPa height anomalies (m)for 2003, with respect tothe 1968-1996 mean.(Courtesy ofB. Chapman)

Figure 5.3(a) Summary of SouthPole total ozone (in DU)and stratospherictemperatures (°C)measured byozonesondes during2003. Three selectedprofiles of altitude vsozone partial pressure(mPa) are shown in thelower panels; (b) prior tothe 2003 ozone hole;(c) the minimum totalozone; and (d)postozone hole.(Courtesy B. Johnson andS. Oltmans,NOAA/CMDL)

hemisphere winter and spring anomalies.In the case of winter, the anomaliesexceeded 4°C in broad areas of NorthAmerica and Russia.

The record low sea-ice extent in 2002was followed by a modest recovery in

2003 to annual values similar to those in2001 (Figures 5.2a, b). Both spring andsummer had large increases over 2002, butfurther sea-ice extent decreases were evi-dent in the northern autumn and winter,with autumn having its lowest extent dur-ing the period of record (Figure 5.5).Surface air temperature anomalies exceed-ed 4°C over the majority of the Arcticbasin during October, and this aided indelaying freeze up across the sea surface,leading to a record-low sea-ice extent forthe month. The rate of refreezingremained low throughout November andDecember, which was reflected in the iceconcentrations and extent. Sea-ice extentanomalies for October also reached theirlowest monthly values since the early1970s (Figure 5.2b). The low October sea-ice was comparable only to 1995, whenthe large negative anomaly (over 1 millionkm2) was sustained for several months,but did not quite reach the same value asfor October 2003 (1.39 million km2).

33

5. Polar climate

Figure 5.5Time series of annualand seasonal sea-iceextent in the northern

hemisphere, 1901-2003(annual values from

Vinnikov, et al., 1999;seasonal values courtesy

of B. Chapman, updatedfrom Chapman and

Walsh, 1993)

6.1 North America6.1.1 Canada

TEMPERATURE

For the period of record that begins in1948, nationally averaged temperatures in2003 ranked as sixth highest. Thiscontinued the recent trend of historicallyhigh temperature rankings in which five ofthe last six national annual averages(1998–2003) fall within the top 10 of allaverages. The exception is 2002, as shownin Figure 6.1. Exceptionally warmconditions were experienced across theCanadian Arctic during 2003 (Figure 6.2), asnoted in subchapter 5.2. In spite of therelatively high national average in 2003,some of the coldest conditions in the past20 years occurred during the winter seasonin parts of eastern Canada, notably in earlyMarch. During that month, the Great Lakeshad their greatest ice cover extent since1996, with the surfaces of Lakes Superior,Erie and Huron covered by 98 per cent. Inthe Gulf of Saint Lawrence, the ice extentwas 50 per cent above normal.

PRECIPITATION

Figure 6.3 shows the precipitationanomalies across Canada in 2003. Thelargest precipitation anomaly occurred inthe Canadian Arctic where regionalconditions were the fourth wettest onrecord. In contrast, drier-than-normalconditions continued in 2003 in the westernprovinces, with parts of the Canadianprairies having experienced their fourthyear of drought. Annually, averagedprecipitation was below normal regionally

from the coast of British Columbia throughthe Prairie Provinces where conditions werethird driest on record. The dry conditionscontributed to one of Canada’s worst fireseasons in the last 50 years in the westernprovinces. Dry conditions also extendedeastward across northern Ontario to parts ofNewfoundland and Labrador.

34

Chapter 6

Regional climate

Figure 6.1Canadian nationwideannual temperaturedepartures (blue line)and long-term trend (redline) for the period1948-2003. Anomalies(in °C) were determinedwith respect to the1951-1980 baseperiod. (Courtesy of theMeteorological Serviceof Canada, EnvironmentCanada)

Figure 6.2Annual temperaturedepartures (in °C) acrossCanada in 2003, withrespect to the 1951-1980 base period.

Figure 6.3Annual precipitationdepartures (in per cent)across Canada in2003, with respect tothe 1951-1980 baseperiod.

SIGNIFICANT EVENTS: HURRICANES ANDWILDFIRES

Numerous notable weather events occurredacross Canada in 2003, and these includedfour tropical storms and hurricanes thatpassed near or made landfall along theAtlantic coast (i.e., Fabian, Isabel, Juan,and Kate; see subchapter 4.2.1 for athorough discussion of the Atlantichurricane season). By far the mostdestructive of these tropical cyclones washurricane Juan, which struck Halifax, NovaScotia, on 29–30 September. Juan formedsouth-east of Bermuda, intensified as itmoved directly northward, and madelandfall at Halifax as a category 2 hurricanewith sustained winds of 158 km h-1 andgusts of 185 km h-1. The strong windsgenerated a storm surge of 1.5 m and raisedthe water level of Halifax harbour to arecord 2.9 m. The storm weakened as itmoved over Nova Scotia, and eventuallypassed over Prince Edward Island as amarginal category 1 hurricane. Juan wasonly the fourth category 2 hurricane tostrike Nova Scotia since the early 1800s,and only the second hurricane to hit PrinceEdward Island since 1930. At its peak, Juanleft more than 300 000 homes withoutpower across the two provinces.

Despite wildfires burning only aroundhalf as many hectares (1.6 million hectares)as the running 10-year average, the impactsfrom the 2003 fire season made it one ofCanada’s worst. The most destructive firesoccurred in British Columbia in August andSeptember, which experienced its worst fireseason as well as the most expensivenatural disaster in the province’s history.According to the Canadian InternationalForest Fire Centre, in total, 2 500 wildfiresburned 6 863 km2 of forests, rangeland andresidential areas across British Columbia,which is 11 times the annual average area

burned in the province over the last 10years. Over the course of the season, morethan 50 000 residents of British Columbiawere evacuated due to fires and relatedsmoke in 2003, which was the secondlargest evacuation in Canadian history.

Approximately one month after thefires had abated, a week-long autumn rainevent in mid-October generated extensiveflooding across British Columbia. The dryrivers and creeks filled quickly with runoffand sediment from the burned areas. TheMeteorological Service of Canada estimatedthat the intense multiday downpour waslikely the largest deluge to strike the WestCoast of Canada in more than 200 years.

6.1.2 United States

TEMPERATURE

Temperatures averaged across theconterminous United States ranked as 14thhighest in the 1895-to-present record. Thelast five five-year periods (1999-2003, 1998-2002, 1997-2001, 1996-2000 and 1995-1999)were the warmest in the last 109 years ofnational records (Figure 6.4). Positivetemperature anomalies were persistent inthe western states while cooler-than-averagetemperatures were recorded along theeastern seaboard. Significantly above-average temperatures dominated theweather west of the Rocky Mountainsduring the summer months (June–August)with both Idaho and Utah having theirwarmest summers on record, averagedstatewide. July 2003 was exceptionallywarm in the west, with Idaho, Wyoming,Utah, Colorado, Arizona and New Mexicoeach reporting their warmest July in thepast 109 years. Idaho and Wyoming alsobroke the statewide records for the monthof August.

PRECIPITATION

Precipitation across the United States in2003 was characterized by moderate yetpersistent deficits in the west, extending thelong-term regional (four–five years)drought, and above average wet anomaliesin the east. At the beginning of the year,approximately one-third of the contiguousnation was in moderate-to-extreme drought,(as defined by the Palmer Drought SeverityIndex (PDSI)). Although improvements

35

6. Regional climate

Figure 6.4Annual temperature

averaged across thecontiguous United States

from 1895 through2003. (Data courtesy of

the U.S. HistoricalClimate Network; Karl,

et al., 1990)

were evident in the summer months, largelydue to mitigating rains in the Plains andGreat Lakes, drought extent again increasedin the fall, with 42 per cent of the 48contiguous states in drought in October(Figure 6.5). At the peak of the 2003drought in late September/early October,approximately 80 per cent of the westernUnited States, from the Rockies to thePacific coast, was affected by moderate toextreme drought. This rivals the greatestdrought extents of the twentieth century inthe west. By the end of 2003, nearly 70 percent of the western United States remainedin moderate to extreme drought, which alsoextended in western Canada and northernMexico.

Several landfalling hurricanes inSeptember and October 2002 also helpedto break the four-year drought for much ofthe south-east. Apart from January 2003,the remainder of the year was mostly wetfor much of the eastern seaboard withVirginia (exceeding the previous annualrecord by the end of November), NorthCarolina and Maryland all breaking theirannual precipitation record. Septemberrainfall along the East Coast was alsoenhanced by the landfall of hurricaneIsabel mid-month along North Carolina’sOuter Banks. Heavy rain fell from theNorth Carolina coast northward throughVirginia and into West Virginia andPennsylvania. Storm-total precipitationexceeding 150 mm was measured in north-western Virginia. Inland flooding wasexacerbated by several months of very wetconditions.

The prolonged cool/wet conditions inthe eastern United States and warmth in thewest were related to several persistentcirculation features, including an anomalousflow of marine air into the western UnitedStates, an enhanced subtropical jet stream,and increased storminess across the southand east. The persistence of thesecirculation features was linked to threedistinct circulation regimes. During October2002–January 2003, they were linked to acombination of El Niño and the negativephase of the Arctic Oscillation (AO). DuringApril–June they were associated withanomalously zonal flow across the country,combined with major Appalachian cold-airdamming events in the east duringApril–May, and a series of major coldfrontal passages in the east during June.

The eastern seaboard had an active2002–2003 snow season beginning in

November 2002. Snowfall totals of 30–60cm fell across much of central New Yorkand north-east Pennsylvania, and into partsof New England on 2–4 January.

Significant snow accumulationsoccurred over parts of the south-east in lateJanuary when 10–30 cm of snow fell acrossparts of North Carolina, including the OuterBanks where snow is uncommon. The“President’s Day Snowstorm” in mid-February had some of the largest snowaccumulations along the mid-Atlantic toNew England coasts since the blizzard of1996 (Halpert and Bell, 1997). As a resultof this and other snow events throughoutthe month of February, several cities alongthe East Coast set new monthly snowfallrecords including: Pittsburgh, Pennsylvania;Baltimore, Maryland; Wilmington,Delaware; and Clarksburg, West Virginia.

In mid-March, a large winter stormimpacted Colorado’s Front Range. Snowfallwas heaviest just west of Denver where upto 220 cm of snow was reported. Denverreceived 80 cm over the period 17–19March. This was the second biggestsnowstorm on record for Denver, and thecity reported its snowiest March on record.Other cities with significant storm totalswere Boulder, Colorado, with 130 cm, FortCollins, Colorado, with up to 78 cm andCheyenne, Wyoming, with over 45 cm.Despite the significant snowfall in March,season-to-date snowfall totals in this regionwere only near or below normal. Never-theless, this event did help alleviate short-term water supply concerns, which was theresult of ongoing drought across theregion.

36

Figure 6.5The PHDI for October2003, illustrating themaximum drought extentin the west in 2003. ThePHDI describes long-termdrought conditions(Heim, 2002).

WILDFIRES

There were three main centres of actionwith respect to wildfire activity in 2003: thesouth-west, especially New Mexico andArizona in May–July, the northern Rockiesin July–September and southern Californiain October. For the season, over 1.5 millionhectares burned in 2003, which is about 80per cent of the 10-year national average.The western United States wildfire seasonbegan later in the year than average as aresult of a wet February–April period. Bythe end of May, a heat wave accompaniedby dry conditions affected much of thewestern United States, which resulted in anincrease in fire activity in the south-west.Persistent hot and dry conditions fuelledfires, which continued well into July. Firedanger indices escalated in early July as aresult of a persistent ridge of high pressuresituated over the western United States. Bylate July, a series of lightning storms led toan increase in fire activity throughout thenorth-west, the eastern Great Basin, and thenorthern Rockies, including parts of GlacierNational Park. Over 280 000 hectares wereconsumed in the northern Rockies wildfireseason and wildfire consumption wasreported to be about 80 per cent of the 10-year average at this time.

Large wildfires broke out in southernCalifornia in late October as a result ofpersistent warm and dry conditions, largeamounts of available fuel in the form ofdead and dormant vegetation, and strongSanta Ana winds. Over 300 000 hectareswere consumed by 15 large fires duringOctober. Higher humidity, coolertemperatures and rainfall in early Novemberhelped firefighters contain these blazes. Themost destructive of these fires, the CedarFire in San Diego County, destroyed over100 000 hectares and was the largest fire inCalifornia since 1932.

6.1.3 Mexico

Precipitation totals averaged over Mexicowere near normal. However, significantrainfall anomalies occurred across thecountry. During the first five months of theyear, the accumulated rainfall across Mexicowas below normal, with the national rainfalltotal being only 76 per cent of the long-term January to May average. In contrast,the latter half of the year was wetter thannormal, with the largest positive rainfallanomalies during autumn (September-November). The national area-weightedmean precipitation for Mexico in 2003 was796 mm, which was 103 per cent of thelong-term mean of 772 mm. Overall, it wasa year of wet and dry contrasts, withsignificant wet anomalies over the isthmusof Tehuantepec and the north-easternstates, as well as in the southern BajaPeninsula, and dry anomalies over centralMexico and the northern Baja Peninsula.

The rainy season, which normallyextends from the end of May to mid-October, began with drier than normalconditions during May acrossapproximately 60 per cent of the country.During the June-August period, monthlytotals were normal to slightly belownormal, with the driest conditionsconcentrated in the states of Chihuahuaand Sinaloa. Heavy precipitation events inSeptember and October across north-western, north-eastern, and central Mexicoled a recovery from deficits experiencedduring the first half of the summer wetseason. Several of these heavy rainfallevents produced flooding over centralMexico (i.e., Guanajuato, Queretaro andHidalgo states). Countrywide, for the entiresummer rainy season (May–October),precipitation averaged 110 per cent ofnormal (Figure 6.6), and September andOctober were 122 per cent and 147 percent of normal, respectively. The broadestarea of below normal precipitation wascentred in the large northern Mexican stateof Chihuahua, where a significant long-term (over two years) moisture deficitcontinued in 2003.

The heavy rainfall events that occurredduring the latter part of the rainy seasonwere related to the influence of tropicalstorms and an active period of tropicaleasterly waves. During the 2003 season,eight tropical cyclones made landfall inMexico on the Pacific and Atlantic coasts,combined. For the period 1970–2003, this

37

6. Regional climate

Figure 6.6May-October 2003

precipitation anomalies(mm) for Mexico, withrespect to the 1941-2002 base period.

(Courtesy of ServicioMeteorológico

Nacional, Mexico)

was second only to 1971 when ninetropical cyclones made landfall in Mexico.The north-eastern Pacific region wasparticularly active for tropical cycloneimpacts, because five tropical cyclonesmade landfall in Mexico. Above-normalrainfall in north-east Mexico was associatedwith two Gulf of Mexico tropical cyclonesthat moved westward in July and August(Claudette and Erika) and eastern northPacific tropical storms that recurved to thenorth-east in October (Nora and Olaf). InBaja California and western Sonora, heavyprecipitation was associated with twohurricanes that moved northward throughthe Gulf of California in August andSeptember (Ignacio and Marty), and twosystems that crossed the mouth of the Gulfof California in October (Nora and Olaf).For the season, seven tropical cyclonescrossed into northern and western Mexicofrom July through early October, andrainfall from these systems helped reversethe drought pattern across northern Mexico(see subchapter 4.2 for a summary of theAtlantic and eastern North Pacific tropicalcyclone seasons). During the five-monthrainy season (June–October), the nationalprecipitation index for Mexico averaged 112per cent of normal, but the majority of thissurplus was attributed to tropical cyclonerainfall rather than the normal, morewidespread monsoon rainfall.

November–December 2003 werecharacterized by a return to drier conditionsand below-normal precipitation across mostof the country. In addition to the dryconditions late in the year, a severe coldoutbreak occurred across north-west Mexicoat the end of December. This cold snapbegan with the arrival of an extremely coldand dry air mass from the north on 27December. By the next morning,temperatures plunged to –11.7°C in Yecora,Sonora, and –11.4°C in Nueva CasasGrandes, Chihuahua. On 29 December,Guachochi, Chihuahua, reported a low of–14.7°C, and Basaseachi, Chihuahua,reported a low of –16.1°C. These record-low temperatures were accompanied bydew-point temperatures of –23°C across theinterior north-west, which illustrates theextremely dry air mass during the event.

6.2 Central AmericaMuch of Central America and the Caribbeancontinued to be relatively dry in 2003(Figure 6.7), extending the overall dry

conditions seen in the region during thepast few years. The drought, which settledinto the region in 2001 and 2002, was amajor contributor to food shortages andmalnutrition, especially in rural areas. Thevast majority of the region recorded below-normal annual precipitation totals, withthose in El Salvador, southern Honduras,and north-western Nicaragua being themost anomalous relative to the 1961–1990base period. However, the largest absoluteprecipitation deficits were observed innorth-eastern Nicaragua and easternHonduras, which are typically among thewettest regions in Central America. Theseareas reported annual precipitation totalsthat were approximately 1 500 mm, or 60per cent, below normal.

Central America generally has abimodal distribution of precipitation, withrelative maxima in May–June andSeptember–October. The largest monthlydeficit on the isthmus emerged during thissecond climatological peak, because theSeptember anomaly accounted for nearly 25per cent of the annual deficit across theregion. Deficits during the wettest part ofthe year (i.e., June–September) contributedto approximately 72 per cent of the annualdeparture. Because this region typicallyreceives such abundant rainfall, heavy rainevents often occur even when monthly andannual precipitation totals are belownormal. In May, over 60 000 people wereaffected by heavy rains and flooding inCosta Rica and Panama as the rainy seasonfirst arrived, while drought conditionscontinued to grip the remainder of theregion.

Lingering tropical systems broughtheavy rainfall and, consequently, landslidesand flooding to El Salvador in October,while heavy rains returned to Costa Ricaand Panama in late November and earlyDecember. According to the Costa RicaNational Meteorological Institute, an entiremonth’s worth of rain (~200 mm) fell in a

38

Figure 6.7Average of standardizedmonthly precipitationanomalies for 2003.Data are from theCAMS-OPI analysis for2003, with anomaliesdetermined from the1961-1990 baseperiod. (Courtesy ofNOAA/CPC). Note thatthe “weighted anomalystandardizedprecipitation” (WASP)index is based solely onmonthly precipitationdata. The index iscomputed using monthlyprecipitation departuresfrom the long-termaverage, which isstandardized by dividingby the standarddeviation of monthlyprecipitation. Thestandardized monthlyanomalies are thenweighted by multiplyingby the fraction of theaverage annualprecipitation for thegiven month. Theseweighted anomalies arethen summed overvarying time periods—here, three, six, nine and12 months. In the figure,the value of the givenWASP index has itselfbeen standardized.

single day across the southern provinces ofthe country. Some areas in Honduras alsoreceived a break from the dry conditions asheavy rains fell along the northern coast inDecember. Precipitation totals for Decemberwere nearly twice their climatologicalaverage in the Honduran departments ofAtlantida and Yoro.

The Caribbean region also experienceda mixed rainy season, with heavyprecipitation occurring amid overall dryconditions. North-west Haiti, which hassuffered under recurrent drought in recentyears, had above-normal spring rainfall, andheavy rains brought flooding to the north-eastern part of the country. Heavy rains andflooding inundated the Dominican Republicin late November and early December aswell. The floodwaters forced over 47 000people out of their homes and swamped200 000 hectares of agricultural land at atime when planting for the next harvestwould have typically been under way.

Surface temperatures were aboveaverage across Central America and theCaribbean during 2003. Annual averagetemperatures were 1.0°–1.5°C above normalin western Honduras, eastern Cuba andsouthern Guatemala. Higher-than-averagetemperatures, which coincided with rainfallin January, contributed to a 30 per cent

increase in dengue fever cases inGuatemala, compared with January 2002(source: Guatemala Ministry of Health;courtesy of the International Society ofInfectious Diseases).

6.3 South AmericaSouth America experienced a variety of wetand dry extremes during 2003 (Figure 6.8).Abnormally dry conditions were prevalentacross the north-eastern half of thecontinent, including southern Chile andcentral-western Argentina. Peru, Ecuador,northern Argentina and Chile were amongthe regions with above-normal annualprecipitation totals in 2003.

The South American continentgenerally receives its largest amounts ofprecipitation across the Amazon River basinwith a secondary maximum stretching tothe south-east. The dry conditions of 2003were most extreme across an area fromsouth-eastern to north-western Brazil andsouth-eastern Colombia (Figure 6.8). North-west and extreme southern Brazil generallyreceive precipitation throughout the year,although most rainfall typically occursduring the austral spring (September,October and November). In 2003, thebelow-normal annual precipitation totalswere due to precipitation deficits thatoccurred throughout the year. In contrast,south-eastern and central Brazil have arainy season that generally extends fromOctober to April. The relatively poor start tothe 2003–2004 rainy season accounted for59 per cent of the annual deficits in thisregion. The exception to these dryconditions came in January, when the SouthAtlantic convergence zone typically has astrong influence on south-eastern Brazil,and heavy rainfall brought flooding andlandslides to the region. The states of Riode Janeiro, Sao Paulo, Espirito Santo andMinas Gerais, in particular, were mostaffected.

The relatively wet conditions in north-western and southern Argentina andportions of north-western South America,particularly in Peru, are notable in theannual standardized anomalies (Figure 6.8).The affected regions in Peru and southernEcuador generally receive precipitationthroughout the year, with the maximumoccurring between January and March. Thisis in contrast to the coastal regions ofEcuador and extreme northern Peru, whichtypically receive less rainfall than the

39

6. Regional climate

Figure 6.8Average of standardized

monthly precipitationanomalies for 2003.

Data are from theCAMS-OPI analysis for2003, with anomalies

determined from the1961-1990 base period

(see Figure 6.7 for adefinition of the WASP

index). (Courtesy ofNOAA/CPC)

summer, although there were some regionalflooding events. Several sunshine recordswere broken in 2003: De Bilt in theNetherlands recorded the sunniest Februaryin its 103-year record, and the UnitedKingdom experienced its sunniest yearsince readily available records began in1961.

Anomalous warmth was widespread inall four seasons, particularly in the summer(see subchapter 6.4.2 on the summer heatwave), with only the three Baltic States andan area around the north-west shore of theBlack Sea colder than the 10th percentile inwinter and spring (see Figures 6.9a, b).Western Europe and parts of eastern Europewere exceptionally warm during thesummer (Figure 6.9c) and in parts of centralEurope it was likely that the summer wasthe warmest since 1540 (Beniston, 2004). InJune, several stations in Switzerlandrecorded temperatures more than 7°Cabove the 1961–1990 average (Figure 6.11),making this an unprecedented event in thehistorical record (Fink, et al., 2004).Although some station records may havebeen affected by urban warming,homogenized records from other parts ofEurope were consistently 5°C warmer thanaverage for several months. Glaciers in theEuropean Alps lost an average thickness of

ice equivalent to approximately 3 m ofwater, nearly twice as much as during theprevious record year of 1998 (1.6 m). TheMediterranean and Near East region(30°–40°N, 20°W–60°E) had their warmestland and SST anomaly on record for bothJune and July. In contrast to the recordheat, regions further east had much coolertemperatures. Moscow experienced one ofits coldest Junes on record, with the firstsnowfall in June since 1963 (see subchapter6.6.3). When averaged across 45°–65°N,25°W–60°E, June temperatures over landareas were close to normal. However, Julyand August were much warmer thanaverage, with land temperature anomaliesof 1.42° and 1.95°C, respectively. The hightemperatures and dry conditionsexacerbated forest fires that burned acrosssouthern France and Portugal in July. Incontrast to the devastating flooding acrosscentral Europe in summer 2002, July sawthe Danube and Elbe Rivers at their lowestlevels for several decades, while Hungaryhad its worst drought since 1950. Therecord heat wave spread across most ofwestern Europe in August. At Paris (Orly),temperatures reached or exceeded 35°Cduring 10 consecutive days, and the UnitedKingdom recorded its highest-ever dailymaximum temperature of 38.5°C atFaversham, Kent*, and its first record oftemperatures above 100°F.

The warmth persisted into autumn(Figure 6.9d), with average autumntemperatures across the western North Searegion exceeding the 98th percentile. Theyear ended with more extremes; Decembertemperatures across Europe were more than2°C above the long-term average, whilesevere regional flooding in south-eastFrance forced 15 000 people from theirhomes.

6.4.2 Summer heat wave

Most of Europe was affected by an extremeheat wave during the boreal summer.

41

6. Regional climate

Figure 6.10Precipitation anomaliesfor the European region

for the period Dec2002-Nov 2003,

expressed as apercentage with respectto 1961-1990 (Source:GPCC, 1998; Rudolf,

et al., 1994)

Figure 6.11June temperature

anomalies (°C) forGeneva, Switzerland.

(Source: MonthlyClimatic Data for the

World)

* The temperature observed at Faversham,

Kent, of 38.5°C is recognized by The Met

Office as the UK record high temperature.

There is currently some discussion about this

value (Burt, 2004; The Met Office, 2004).

However, the previous record was also

broken at other sites, including a 38.1°C at

Kew, Royal Botanic Gardens, west London.

According to the World Health Organizationand Reuters/Associated Press, in France,11 000 heat-related deaths were reportedbetween late July and mid-August. Thesedeaths resulted not only from the extremedaily heat, but from the frequency ofextremely hot and dry days in areasunaccustomed to such conditions. Theseconditions were part of a prolonged warmand dry spell that began in April inresponse to a persistent upper-level ridge ofhigh pressure centred over the continent.

Most of Europe (southern Spain tocentral France) experienced daily maximumtemperatures exceeding 34°C for 30–50days during JJA, thus, 20 days more thanaverage (Figure 6.12). In central andnorthern France there were 10–30 days withmaximum temperatures above 34°C, wellabove the mean of five days or less. Twodistinct periods of exceptional heatoccurred during the season—the first inJune and the second during the first half ofAugust. The August heat wave was themost serious, because it coincided with thenormal peak in summer temperatures andwas accompanied by an almost completeabsence of rainfall. Daily maximumtemperatures during this period averagedmore than 40°C across most of inlandSpain, 36°–38°C across southern and centralFrance, and 32°–36°C across northernFrance (Figure 6.13, contours).

In general, these temperatures were7.5°–12.5°C above average (Figure 6.13,shading). Over Germany, both June andAugust were the warmest months since thebeginning of the twentieth century. Overallthe summer (JJA) average was the hottestfor Germany since 1901 (averagetemperature of 19.6°C; 3.4°C above the1961–1990 base period), and, with theexception of some northern and north-western stations, it was the hottest Germansummer since the beginning of recordedmeasurements. The highest minimumtemperature (27.6°C) ever recorded inGermany was observed on 13 August inWeinbiet, in the middle Rhine valley (Fink,et al., 2004).

Surface temperatures were well aboveaverage across Europe throughout the Aprilto August period, with anomalies in excessof +2.5°C across central Europe and +1.5°Cacross northern Europe (Figure 6.14a,shading). Precipitation totals were also wellbelow average during this period, withdeficits of 75–100 mm observed throughoutcentral Europe (Figure 6.14b, shading). The

largest temperature and precipitationanomalies coincided with the mean positionof a very persistent upper-level ridge(Figure 6.15a). An enhanced North Atlanticjet stream was also located just to the northof these regions (Figure 6.15b), resulting inlarge-scale sinking motion and a reductionin the number and intensity of convectivestorms, precipitation events and cold frontalpassages.

These conditions were associated witha larger-scale anomalous circulationcharacterized by below-average heights athigh latitudes from Canada to Great Britain,and above-average heights in the middlelatitudes from the north-eastern UnitedStates to eastern Europe, as well as in the

42

Figure 6.12Anomalous number ofdays with maximumsurface temperaturesreaching 34°C for JJA2003. Anomalies aredepartures from the1971-2000 base perioddaily means.

Figure 6.13Mean daily maximumsurface temperature(contours, interval is 2°C)overlaid with (a) surfacetemperatures anomalies(°C); and (b)precipitation anomalies(mm) for April-Aug 2003.Anomalies are departuresfrom the 1971-2000base period monthlymeans.

Figure 6.14Mean 500-hPa heights(contours, 60 m interval)overlaid with (a) surfacetemperatures anomalies(°C); and (b)precipitation anomalies(mm) for April-Aug 2003.Anomalies are departuresfrom the 1971-2000base period monthlymeans.

subtropics (Figure 6.15). This anomalypattern (Figure 6.16) reflected a strongpositive phase of the eastern Atlanticteleconnection pattern (Barnston andLivezey, 1987). The standardized easternAtlantic index values for June–August were2.1, 0.9 and 1.3, respectively.

6.5 Africa6.5.1 North Africa

The Mediterranean coast of North Africareceives most of its annual rainfall duringthe period from October to April, largelyfrom mid-latitude cyclones and theirassociated cold frontal incursions intosubtropical regions. During most of the1980s and 1990s there were severe droughtepisodes across northern Africa. During thelast three years, a succession of wetautumns (September to November) hasbrought more rainfall to the northern partof the region. These above-average rainscontinued during October 2002–April 2003,with totals exceeding 300 mm across thenorthern half of Morocco and Tunisia, andnorth-east Algeria (Figure 6.17a). Theseamounts were generally 125 mm or moreabove the long-term average (Figure 6.17b).

The heaviest precipitation occurredduring January across northern Algeria andTunisia. Precipitation totals of 25–100 mmduring 22–28 January generated some ofthe worst flooding observed during the last10 years in northern and central Tunisia. InMorocco, anomalously wet conditionsprevailed during January–May.

With the exception of the highelevations of the Atlas Mountains, summersare generally dry and hot. DuringJuly–August 2003, monthly temperatures insome locations in North Africa averaged3°–4°C above normal. In Morocco, newmonthly temperature records were set inseveral areas during the period. The heatwas most extreme during August whenseveral northern Moroccan citiesexperienced new all-time maximumtemperature records, including Rabat(44.6°C), Kenitra (47.7°C) and Tangier(43.5°C). These record maximums occurredduring the same period as the Europeanheat wave (see subchapter 6.4.2).

Several locations in North Africa wereaffected by severe storms and heavyprecipitation during the latter part of theyear. Much of Morocco and northernAlgeria received above-normal rainfallduring October. In some instances, the wetconditions were a result of short-durationheavy episodic rainfall events. CentralMorocco recorded 85 mm of rainfall in lessthan 24 hours during 21–22 October whichaccounted for nearly 85 per cent of theannual total of 102 mm. Another episodicevent occurred on 17–18 December in

43

6. Regional climate

Figure 6.15Mean 200-hPa (a)

heights and anomalies(m); and (b) wind speedsand anomalies (m s-1) for

April-Aug 2003.Anomalies (shading) are

departures from the1971-2000 base period

monthly means.

Figure 6.16The positive phase of the

eastern Atlanticteleconnection pattern,

as indicated by 500-hPaheight anomalies

(positive heightanomalies are shaded

orange, negative heightanomalies are shaded

blue).

Figure 6.17Oct 2002-April 2003:(top) total precipitation

(mm) and (bottom)anomalies for the

northern Africa.Anomalies are departures

from the 1979-1995base period monthly

means. (CAMS-OPI datacourtesy of NOAA/CPC)

44

northern Morocco producing 127 mm ofrain in 24 hours. This total was 44 per centof the average annual rainfall and 400 percent of the December monthly average.

6.5.2 West Africa

Rainfall

West Africa can be divided into two quasi-homogeneous regions based on rainfall.These regions are the African Sahel and theGulf of Guinea. The African Sahel is theregion between 10°–20°N, 18°W–20°E(Figure 6.18, boxed region) and receivesapproximately 90 per cent of its meanannual rainfall during June–September. Therainy season is monsoonal in character andis closely related to the north–southmovement of the Inter-TropicalConvergence Zone (ITCZ), which reachesits northernmost position in August.Seasonal precipitation exhibits a strongmeridional gradient, with average totalsexceeding 600 mm in the south andreaching 100–300 mm in the north.

From June to September 2003,precipitation totals exceeded 100 mm aboveaverage across much of the central Sahel(Figure 6.18b). Consequently, the 2003rainy season was the second wettest since1990 (Figure 6.19). In contrast, rainfallalong the Guinea coast was below averageduring the period, with exceptionally dryconditions observed in July and August.This rainfall pattern reflected an amplifiedmonsoon circulation, with the ITCZremaining consistently north of itsclimatological position from May untilOctober. The enhanced monsoonal flowwas evident even up to 850 hPa (Figure6.20) and contributed to a deep penetrationof moist, unstable air well into the Sahelregion. At 925 hPa, strong south-westerlywinds averaging 6–9 m s-1 extended wellnorthward into the Sahel region (Figure6.21). At upper levels, the enhancedmonsoon circulation was associated with astrengthening of the subtropical ridgesacross both hemispheres (see Figure 4.7),and with a corresponding amplification ofthe tropical easterly jet. These sameconditions contributed to the above-normal2003 Atlantic hurricane season (seesubchapter 4.2.1).

The inverse relationship betweenrainfall anomalies in the Sahel and Guineacoast regions has been a recurring pattern of

West African climate variability (Janicot,1992; Ward, 1998). From the 1970s to themid-1990s the prevailing pattern was a drySahel and a wet Guinea coast. Since 1995,rainfall has returned to near-normal levelsacross the African Sahel, in association withan overall strengthening of the West Africanmonsoon circulation. Another notable aspectof these stronger monsoon circulations hasbeen a significant increase in Atlantichurricane activity since 1995 (see subchapter4.2.1).

During July–September 2003, positiveSST anomalies in the Atlantic persistedalong the equator and southward along thewest coast of southern Africa. This

Figure 6.18June-Sep 2003: (top)total precipitation (mm)and (bottom) anomaliesfor the African Sahel andGulf of Guinea regions.The boxed area denotesthe approximateboundaries of the Sahelregion. Anomalies aredepartures from the1979-1995 periodmonthly means. (CAMS-OPI data courtesy ofNOAA/CPC)

Figure 6.19Precipitation index timeseries for the period1951-2003 expressedin terms of percentiles.The index wascalculated from the1971-2000 base periodseasonal means for theAfrican Sahel duringJune-Sep using a gammadistribution.

Tanzania, southern Sudan and southernEthiopia) experiences a bimodal rainfallregime with its first rainy season (termed“long rains”) occurring during March–Mayas the rainbelt shifts northward from thesouthern hemisphere. It then experiencesits second rainy season (termed “shortrains”) during October–December as therainbelt shifts southward from the AfricanSahel and Sudan regions. Areas close tolarge water bodies, such as Lake Victoriaand parts of the coastal strip, receivesubstantial rainfall throughout the year. Inthe northern sector (comprising the rest ofSudan and Ethiopia, northern Somalia,Djibouti and Eritrea), rainfall normallyoccurs from May–June toSeptember–October.

SOUTHERN SECTOR

Widespread dry conditions characterizedthe sector during January, February andApril 2003 (Figure 6.22). However,significant amounts of rainfall were receivedover vast areas of the southern sector inMarch 2003. Despite these rains, the annualtotal rainfall for the region was generally300 to 500 mm below average (Figure6.23), which led to drought conditions andfood shortages for close to two millionpeople.

Generally dry conditions characterizedthe sector during the months of April toSeptember 2003.

EQUATORIAL SECTOR

January to mid-March is a climatologicallyhot and dry period over the equatorialsector. The 2003 season of “long rains”started in earnest in mid-March but a longunprecedented dry spell engulfed the sectorin early April. It was not until the secondhalf of April that significant rainfall wasobserved across the sector. Some locationsin the marginal areas were characterized byshort-lived, intense precipitation events(Figure 6.24). The wet anomalies oversouthern Ethiopia, central Tanzania andparts of western and central Kenya resultedin some locations experiencing their wettestconditions on record (since 1961). Theheavy rains continued into May 2003,producing near-record rainfall totals andflooding in northern Kenya, southernEthiopia and southern Somalia. Enhancedprecipitation was observed over mostlocations of the western highlands of Kenyafrom June to August. DuringOctober–December, below normal rainfall

was observed in southern Ethiopia, Kenyaand southern Somalia.

NORTHERN SECTOR

Much of the Greater Horn of Africa hassuffered from persistent drought conditionsfor several consecutive years. This has beenparticularly true in the northern sector,where rainfall deficits during 2002 createdfood shortages for nearly 14 million peoplein Ethiopia and Eritrea by early 2003.

46

Figure 6.22Jan-Mar 2003 per centof normal precipitationover the Greater Horn ofAfrica.

Figure 6.23Annual precipitationanomalies for 2003across the Greater Hornof Africa. Anomalies aredepartures from the1979-1995 baseperiod. (CAMS-OPI datacourtesy ofNOAA/CPC)

Figure 6.24Cumulative rainfall atWajir, Kenya (located innorth-eastern Kenya),during 2003 (red),compared with the long-term annual mean (blue).

The onset of the rainfall season in 2003occurred during the month of June, whichwas somewhat later than normal for thesector. Rainfall totals were near normal inJune, but declined in July and August overthe southern parts of the sector, while thecentral and western parts of the sectorreceived near-normal rainfall during thisperiod. Rainfall in excess of 200 mm overthree consecutive months (June–August)resulted in normal to above-normalconditions. Parts of central Ethiopia, as wellas eastern and southern Sudan, continued toreceive above-average rainfall in September,which led to one of the best harvests of thelast five years. However, severe droughtconditions continued in northern Somaliaand areas of southern and eastern Ethiopia.

6.5.4 Southern Africa

RAINFALL

The rainy season over much of southernAfrica extends from October to April, withthe largest totals typically observed betweenDecember and March. Exceptions to thisunimodal rainfall distribution are found inthe northern areas where rainfall occursduring October–December and March–Mayin response to the north–south movementof the ITCZ; the south-western Cape regionof South Africa where rainfall occurs during

both October–April and June–August; andalong the southern coast of South Africaand extreme northern Tanzania borderingLake Victoria, where rainfall is typicallyobserved throughout the year.

October to April rainfall in southernAfrica is normally suppressed during ElNiño years and enhanced during La Niñayears (Ropelewski and Halpert 1987, 1989,and 1996; Hastenrath, Greischar and VanHeerden, 1995; Dai, Fung and Del Genio,1997; and Thiaw, Barnston and Kumar,1999). The October–April 2002–2003 rainyseason was below average, with rainfalldeficits observed throughout the region(Figure 6.25b). Precipitation wasconsistently below average from October2002 to February 2003, and again in April,with March being the only month withabove-average totals for the region as awhole (Figure 6.25d). This indicated asignificant delay in the onset of the Octoberrains and an early return to dry conditionsin April. These below-average rainfallanomalies were consistent with moderate ElNiño conditions observed during theperiod.

The most significant rainfall deficitswere found in the east across the normallyheavy precipitation areas of eastern SouthAfrica, Botswana, Zimbabwe and thesouthern half of Mozambique. Precipitationtotals in these areas (Figure 6.25a) averagedonly in the fifth to twentieth percentile ofoccurrences (Figure 6.25c), which led to acontinuation of drought conditions fromeastern and northern Botswana into westernZimbabwe, northern South Africa, southernMozambique and Swaziland.

The south-western Cape region ofSouth Africa also recorded below-averagerains during October 2002–April 2003,which followed a drier-than-average winterrainy season during June–August 2003.These continuing rainfall deficits weredevastating to agricultural activities in thisarea.

ATMOSPHERIC CIRCULATION

The low-level atmospheric circulationduring the below-average October–April2002-2003 southern Africa rainy seasonfeatured easterly winds along theequatorward flank of the Mascarene Islandshigh pressure system (located over thesouthern Indian Ocean). The easterly windsaveraged 4–8 m s-1 from the east-centralIndian Ocean westward into Madagascarand portions of interior southern Africa

47

6. Regional climate

Figure 6.25Nov 2002-April 2003

southern Africanprecipitation (a) totals

(mm); (b) anomalies(mm); (c) percentiles,based on a gammadistribution fit to the

1971-2000 baseperiod; and (d) monthly

time series ofprecipitation percentiles,

based on precipitationtotals averaged over theboxed region in southern

Africa depicted inpanels (a)-(c).

remaining below normal across most ofAfghanistan and Iran (Figure 6.28). SouthernAfghanistan and north-western Iran receivedonly about 50 per cent of their averageannual precipitation, with deficits exceeding200 mm in some areas. However, elsewherein Afghanistan the drought eased asprecipitation totals generally exceeded thoserecorded during the previous year (Figure6.29). Although seasonal totals remainedbelow average, parts of southern Afghanistanreceived 80 per cent more precipitationduring January and February 2003 thanduring all of 2002. Spring rains also broughtlocalized flooding to the northern provinceof Kunduz, while monsoon rains duringAugust ruptured dams and killed at least 24people in the south-east.

In contrast to the relatively dryconditions across the south, the northernhalf of south-west Asia received above-normal precipitation during 2003 (Figure6.28). South-central Kazakhstan received

nearly three times, or over 200 mm more,than its annual average. Above-normalprecipitation was also observed in easternTurkmenistan, eastern and northernUzbekistan, northern and central Kyrgyzstan,and north-eastern Tajikistan. Annualprecipitation totals were 100–180 mm higherthan the previous year (Figure 6.29) overthese countries. Above-normal precipitationduring the spring and early summer, andheavy mountain snows in November, allcontributed to the above-average annualtotals in northern south-west Asia. Heavyrains and snowmelt in April brought floodingand landslides to areas in Kyrgyzstan andTajikistan, which contributed to Kyrgyzstanrecording eight times as many landslidesduring the entire year as in 2002.

6.6.3 Russia (including EuropeanRussia)

Anomalously warm temperatures wereobserved across the country as a whole in2003, with an annual average surfacetemperature of 1.2°C above the long-termmean, thus continuing the trend of above-average annual temperatures.

During January and February 2003,positive air temperature anomalies prevailedover almost the entire country. Over theCola Peninsula, the monthly averaged airtemperature anomaly for February exceededthe normal by 6 to 7°C. In the first half ofspring, cooler-than-normal weather andabundant precipitation persisted acrossEuropean Russia. In the Volgograd region,ice jams formed on rivers due to sharptemperature rises. Vigorous spring floodingalso swelled small rivers and tributaries inthe Rostov region. In late spring, much ofthe country experienced above-normal airtemperatures, which, in conjunction withbelow-normal rainfall, produced anincreased threat of wildfires in the Irkutskregion and Transbaikalia. Extremely dryconditions in the catchments of the upperand middle Amur River resulted in thelowest water levels in these basins in the100-year instrumental record.

June in European Russia was one of itscoldest in the past 100 years. Monthly meantemperatures were 3-4°C below normalacross the region, and these coldtemperatures were accompanied by above-normal precipitation. The Ural,Volga–Vyatka, and Volga regions, as well asthe northern part of the central region,

49

6. Regional climate

Figure 6.28Annual precipitation

anomalies (in mm) for2003, with anomalies

determined from the1979-1995 base

period. (Data are fromthe CAMS-OPI analysisfor 2003. Courtesy of

NOAA/CPC)

Figure 6.29Difference between the

2003 and 2002 annualprecipitation anomalies

(mm). Positive valuesindicate greater

precipitation totals during2003, with anomalies

determined from the1979-1995 base

period. (Data are fromCASM-OPI (Courtesy of

NOAA/CPC)

accumulated over 200 per cent of theirmean monthly precipitation. In contrast,western Siberia was anomalously warm inJune, with average monthly temperatureanomalies exceeding 5°C. The republic ofKomi and the Nenets Autonomous Okrugarea reported mean monthly temperatureanomalies of 4–5°C above normal. August2003 was one of the warmest over Russia inthe last 100 years (Figure 6.30), as well asbeing one of the wettest months on recordin European Russia. At many stations, notonly monthly, but also daily precipitationrecords were broken.

From October to December, recordwarmth was again recorded throughoutRussia. Record monthly temperatureanomalies were observed during October inthe north of the Magadan region (greaterthan 8°C) and in the middle Ural (3° to4.5°C). In contrast, November in Siberiasaw record cold anomalies of -6°C. InDecember, record warmth was againrecorded throughout Russia with severalstations measuring their highest monthlyaveraged surface temperatures sinceinstrumental records began.

6.6.4 Monsoon

SOUTH ASIAN MONSOON

A notable feature of the south Asiansummer monsoon rainfall in 2003 was itsequitable distribution in space and time,devoid of any prolonged dry spell.

The summer monsoon arrived overKerala, India, on 8 June, seven days afterthe normal arrival date, but then rapidlycovered all of South Asia by 5 July,approximately 10 days ahead of thenormal date (Figure 6.31). A peculiaraspect of the onset phase during 2003 wasthat it occurred in north-east India (5 June)prior to Kerala (though this is notunprecedented, with similar occurrences in1972, 1995, 1996, 1998 and 2000).

50

Figure 6.30Map: surface temperature anomalies (°C) during Aug 2003 across Russia. Inset:precipitation anomalies over European Russia, with two example graphs illustrating therecord wet conditions at two sites: Petrozavodsk and Pskov.

Figure 6.31Isochrones of the onset and advance of the summer

monsoon in 2003 (solid), compared with the long-termmean dates of advance (dashed).

edge of the summer monsoon advancedover northern China. The summermonsoon began its withdrawal aroundmid-August, and withdrew from the SouthChina Sea in the fourth pentad ofSeptember, somewhat earlier than normal(National Climate Center/CMA, 2003).However Viet Nam and Thailand sawheavy rains and severe flooding inOctober, which killed more than 100people. Indonesia suffered from landslidestriggered by heavy rains in January, whilelandslides later in the year in Indonesia,Malaysia and the Philippines caused deathsand disrupted local infrastructure.

6.7 Australasia and theSouth-West Pacific

6.7.1 Australia

The Australian climate of 2003 was largelyinfluenced by the transition from El Niño toneutral conditions in the equatorial PacificOcean during the first half of the year. Thefailure of a La Niña pattern to emerge,which is associated with wetter than normalconditions for much of eastern Australia,meant that the recovery from the severe2002–2003 Australian drought was slow todevelop and spatially inhomogeneous. Thehydrological, agricultural and societalimpacts of one of Australia’s worst droughtsin its recorded meteorological historyremained for much, if not all, of 2003.

52

Figure 6.34Variation of pentad dewpoint temperature (Td)and potential pseudo-equivalent temperaturealong 115°E at the 850-hPa level from the firstpentad of April 2003 tothe second pentad ofOctober 2003.

Figure 6.35Rainfall decilesdetermined for the period1 Jan-31 Dec 2003.

Figure 6.36Australia-wide annual mean precipitation (blue bars), along with the maximum (red line)and minimum (green line) temperature anomalies (°C) plotted with respect to the 1961-90 base period. The 1961-1990 climatological normal rainfall for Australia is472 mm.

6.7.2 South-West Pacific

The 2002-2003 El Niño event had anoticeable impact on the South-WestPacific precipitation anomalies at thebeginning of the year, with above-averagerainfall over Kiribati and below-averagerainfall over much of New Caledonia andFiji. However, as the equatorial PacificOcean settled into a near-neutral ENSOstate, rainfall patterns for the regiongenerally approached normal. Exceptionswere in the Coral Sea, where above-average surface pressures led tosuppressed convection over the southernparts of the Solomon Islands, northernVanuatu and areas toward the Date Line,and a further region of suppressedconvective activity along the Equator fromthe Date Line eastward toward SouthAmerica, including eastern Kiribati and theMarquesas Islands (Figure 6.38). Theseareas of suppressed convection resulted in2003 rainfall totals at Willis Island (358 mm:32 per cent average), Ono-i-lau in Fiji(1 068 mm: 65 per cent average) and HivaHoa in the Marquesas Islands (751 mm: 62per cent average) being among the lowestannual rainfall totals on record.

In some contrast, areas of enhancedconvection affected the region north ofPapua New Guinea, the Marshall Islands,and parts of southern French Polynesiaduring 2003. For the year, the South PacificConvergence Zone was generally farthernorth and east than usual. However, itsmain region of convective activity wasconcentrated in areas east of the Date Linefrom February to May, moving to areaswest of the Date Line from September toNovember. This anomalous position of theSouth Pacific Convergence Zone generatedabove-average annual precipitation in partsof New Caledonia, Tonga and the southernCook Islands.

SSTs and, hence, mean air temperatureswere at least 0.5°C above averagethroughout much of the tropical South-WestPacific in 2003 (Figure 6.39). The largestpositive anomalies occurred along theEquator, from Nauru to eastern Kiribati, withSSTs around 1.0°C above average. The yearwas also one of the warmest years onrecord for mean air temperatures in FrenchPolynesia, with the Marquesas Islands 0.9°Cabove average (at Hiva Hoa) and theTuamotu Islands recording temperatures0.7°C above average at Takaroa. Fiji wasfairly typical of the western half of the

basin, with an annual mean air temperatureof 0.3°C above normal.

In total, the 2002-2003 tropical cycloneseason brought 11 cyclones to the South-West Pacific, above the long-term averageof nine, while the 2003-2004 seasonbrought only four (see subchapter 4.2.2).Two tropical cyclones developed very latein the 2002-2003 season (June). In this case,a Madden–Julian Oscillation event had justpreviously propagated over or near arelatively small region of warm SSTanomalies in the western South Pacific,creating a set of antecedent conditions thatwere uniquely primed for such late tropicalcyclogenesis events.

6.7.3 New Zealand

Anticyclones dominated New Zealand’sclimate in 2003 (Figure 6.40), resulting inthe driest year on record in some southernareas, with precipitation totals only about60 per cent of normal. Rainfall was alsowell below average in north Canterburyand central Marlborough, with totals lessthan 75 per cent of normal. In contrast, the

54

Figure 6.39Annually-averaged SSTanomalies (°C) for2003 across the south-west Pacific basin (lightgreen/orange shadingdenotes warmer-than-average SSTs, and light-blue/blue shadingdenotes cooler than-average SSTs).

Figure 6.38Annually-averaged

outgoing long-waveradiation anomalies(W m-2) for 2003,

represented by shadedareas, and annual

precipitation percentageof average, shown bynumbers. Higher OLR

values (yellow or orange)were associated with

clearer skies and lowerrainfall, while lower OLR

values (blue) wereassociated with cloudy

conditions and typicallyhigher average

precipitation.

North Island experienced the majority ofthe 20 heavy rainfall events to affect thecountry, of which nine produced floods.The year was also exceptionally sunnyover much of the South Island. Nelsonrecorded 2 707 sunshine hours, its secondhighest annual total on record whileWellington, on the southern end of theNorth Island, recorded its sunniest year on

record. The reduced cloud contributed to aNew Zealand-wide average annualtemperature of 12.7°C, 0.1°C above the1971–2000 mean.

June was the warmest on record(records date back over 150 years), withnationwide temperatures 2.0°C aboveaverage. Conversely, July was anomalouslycold. The first week of July broughtsnowfalls of up to 30 cm (~12 in) in theeastern South Island and the North Islandhigh country. This snowstorm wasdescribed as one of the worst in 50 years,and was one of five significant snowfallevents of the winter (Figure 6.41).

The latter part of the year wasdominated by variable west-to-south-westerly flow, which produced severe dryconditions and soil moisture deficitsthroughout Otago, Canterbury and centralMarlborough, and significant anomalies inWairarapa, Nelson, and parts of Northland.North-westerly gales (up to 176 km h-1 atSouth West Cape) affected central andeastern New Zealand in September, alongwith several destructive tornadoes on theWest Coast on 18 September.

55

6. Regional climate

Figure 6.40Mean sea level pressureanomaly map for 2003

around New Zealand(hPa).

Figure 6.41Snowstorms in early July2003 were some of the

heaviest to hit NewZealand in 50 years.

(Image from 11 July2003, courtesy Jacques

Descloitres, MODISRapid Response Team at

NASA GSFC)

56

Chapter 7

Seasonal summaries

Figure 7.1Dec 2002-Feb 2003 (top, °C) surface temperatureanomalies and (bottom) precipitation percentiles basedon a gamma distribution fit to the 1979-2000 baseperiod. Temperature anomalies (1971-2000 baseperiod) were based on station data over land and seasurface temperature over water. Precipitation data wereobtained from a combination of raingauge observationsand satellite-derived precipitation estimates (Janowiakand Xie, 1999). The analysis was omitted in data-sparse regions (white areas).

Figure 7.2Dec 2002-Feb 2003

(top) northern hemisphereand (bottom) southernhemisphere 500-hPageopotential heights(contour interval is 9

decametres) andanomalies (shading).

Anomalies are departuresfrom the 1979-2000base period means.

57

7. Seasonal summaries

Figure 7.3Mar-May 2003 (top, °C)surface temperatureanomalies and (bottom)precipitation percentilesbased on a gammadistribution fit to the1979-2000 baseperiod. Temperatureanomalies (1971-2000base period) were basedon station data over landand sea surfacetemperature over water.Precipitation data wereobtained from acombination ofraingauge observationsand satellite-derivedprecipitation estimates(Janowiak and Xie,1999). The analysis wasomitted in data-sparseregions (white areas)

Figure 7.4Mar-May 2003 (top)

northern hemisphere and(bottom) southern

hemisphere 500-hPageopotential heights(contour interval is 9


Anomalies aredepartures from the

1979-2000 base periodmeans.

58

Figure 7.5June-Aug 2003 (top, °C)

surface temperatureanomalies and (bottom)precipitation percentiles

based on a gammadistribution fit to the

1979-2000 baseperiod. Temperature

anomalies (1971-2000base period) were

based on station dataover land and sea

surface temperature overwater. Precipitation data

were obtained from acombination of

raingauge observationsand satellite-derived

precipitation estimates(Janowiak and Xie,

1999). The analysiswas omitted in data-sparse regions (white

areas).

Figure 7.6June-Aug 2003 (top)northern hemisphere and(bottom) southernhemisphere 500-hPageopotential heights(contour interval is 9decametres) andanomalies (shading).Anomalies aredepartures from the1979-2000 baseperiod means.

59

7. Seasonal summaries

Figure 7.7Sep-Nov 2003 (top, °C)surface temperatureanomalies and (bottom)precipitation percentilesbased on a gammadistribution fit to the1979-2000 baseperiod. Temperatureanomalies (1971-2000base period) were basedon station data over landand sea surfacetemperature over water.Precipitation data wereobtained from acombination ofraingauge observationsand satellite-derivedprecipitation estimates(Janowiak and Xie,1999). The analysis wasomitted in data-sparseregions (white areas)

Figure 7.8Sep-Nov 2003 (top)northern hemisphere

and (bottom) southernhemisphere 500-hPageopotential heights(contour interval is 9


Anomalies aredepartures from the1979-2000 base

period means.

60

Barnston, A.G. and R.E. Livezey, 1987:Classification, seasonality andpersistence of low-frequencyatmospheric circulation patterns.Monthly Weather Review, Vol. 115,pp. 1083–1126.

Behringer, D.W., M.J. and A. Leetmaa, 1998:An improved coupled model for ENSOprediction and implications for oceaninitialization. Part I: The ocean dataassimilation system. Monthly WeatherReview, Vol. 126, pp. 1013-1021.

Bell, G.D., M.S. Halpert, R.S. Schnell,R.W. Higgins, J. Lawrimore,V.E. Kousky, R. Tinker, W. Thiaw,M.Chelliah and A. Artusa, 2000: Climateassessment for 1999. Bulletin of theAmerican Meteorological Society,Vol. 81, pp. S1–S50.

Beniston, M., 2004: The 2003 heat wave inEurope: A shape of things to come? Ananalysis based on Swiss climatologicaldata and model simulations.Geophysical Research Letters, Vol. 31,L02204.

Bruhwiler, L.M., E.S. Kasischke, E.J.Dlugokencky and P. Tans, 2000: Borealbiomass burning during 1998 andanomalous northern hemispheric CO.EOS, Transactions, AmericanGeophysical Union, Vol. 81, 260 pp.

Bureau of Meteorology, 2003: The weatherof 2002-03, Australian Government.Bureau of Meteorology Annual Report2002-03, pp. 237-252.

Burt, S., 2004: The August 2003 heatwave inthe United Kingdom: Part 1—Maximumtemperatures and historical precedents.Weather, Vol. 59, pp. 199-208.

Chapman, W.L. and J.E. Walsh, 1993: Recentvariations of sea ice and airtemperature in high latitudes. Bulletinof the American MeteorologicalSociety, Vol. 74, pp. 2–16.

Chelliah, M. and G.D. Bell, 2004: Tropicalmultidecadal and interannual climatevariations in the NCEP–NCAR

Reanalysis. Journal of Climate, Vol. 17,pp. 1777–1803.

Christy, J.R., R.W. Spencer, W.B. Norris,W.D. Braswell and D.E. Parker, 2003:Error estimates of version 5.0 ofMSU/AMSU bulk atmospherictemperatures. Journal of Atmosphericand Oceanic Technology, Vol. 20,pp. 613–629.

Ciais, P., P.P. Tans, M. Trolier, J.W.C. Whiteand R.J. Francey, 1995: A largenorthern hemisphere terrestrial CO2

sink indicated by the 13C/12C ratio ofatmospheric CO2. Science, Vol. 269,pp. 1098–1102.

Conway, T.J., P.P. Tans, L.S. Waterman, K.W.Thoning, D.R. Kitzis, K.A. Masarie andN. Zhang, 1994: Evidence forinterannual variability of the carboncycle from the NOAA CMDL global airsampling network. Journal ofGeophysical Research, Vol. 99,pp. 22831–22855.

Dai, A., I.Y. Fung and A.D. Del Genio,1997: Surface observed global landprecipitation variations during 1900–88.Journal of Climate, Vol. 10,pp. 2943–2961.

Daniel, J.S. and S. Solomon, 1998: On theclimate forcing of carbon monoxide.Journal of Geophysical Research,Vol. 103, pp. 13249–13260.

Daniel, J.S., S. Solomon, R.W. Portmann andR.R. Garcia, 1999: Stratospheric ozonedestruction: The importance ofbromine relative to chlorine. Journal ofGeophysical Research, Vol. 104,pp. 23871–23880.

Dlugokencky, E.J., S. Houweling,L. Bruhwiler, K.A. Masarie, P.M. Lang,J.B. Miller and P.P. Tans, 2003:Atmospheric methane levels off:Temporary pause or new steady-state?Geophysical Research Letters, Vol. 30,No. 19, 1992.

Fink, A.H., T. Bruecher, A. Krueger, G.C.Leckebusch, J.G. Pinto, U. Ulbrich,

References

2004: The 2003 European summerheatwaves and drought: synopticdiagnosis and impacts. Weather,Vol. 59, pp. 209-216.

Folland, C.K. and T.R. Karl, 2001: Observedclimate variability and change. ClimateChange 2001: The Scientific Basis (J.T.Houghton, et al., eds.), CambridgeUniversity Press, pp. 99–181.

Fu, Q., C.M. Johanson, S.G. Warren,D.J. Seidel, 2004: Contribution ofstratospheric cooling to satellite-inferred tropospheric temperaturetrends. Nature, Vol. 429, pp. 55-58.

Goldenberg, S.B., and L.J. Shapiro, 1996:Physical mechanisms for theassociation of El Niño and west Africanrainfall with Atlantic major hurricanes.Journal of Climate, Vol. 9,pp. 1169–1187.

GPCC, 1998: The Global PrecipitationClimatology Centre [available online athttp://gpcc.dwd.de].

Hall, B.D., J.W. Elkin, J.H. Butler, S.A.Montzka, T.M. Thompson, L. delNegro, G.S. Dutton, D.F. Hurst, D.B.King, E.S. Kline, L. Lock, D.MacTaggart, D. Mondeel, F.L. Moore,J.D. Nance, E.A. Ray and P.A.Pomashkin, 2002: Halocarbons andother atmospheric trace species.Climate Monitoring and DiagnosticsLaboratory Summary Report No. 25,1998–1999 (D.B. King, R.C. Schnelland R.M. Rosson, eds.), NOAA Oceanicand Atmospheric Research, 154 pp.

Halpert, M.S. and G.D. Bell, 1997: ClimateAssessment for 1996. Bulletin of theAmerican Meteorological Society,Vol. 78, pp. S1–S49.

Halpert, M.S. and G.D. Bell, 2003: ENSOand the tropical Pacific. State of theclimate in 2002 (A.M. Waple and J.H.Lawrimore, eds.). Bulletin of theAmerican Meteorological Society,Vol. 84, pp. S1–S68.

Hastenrath, S., L. Greischar and J. VanHeerden, 1995: Prediction of thesummer rainfall over South Africa.Journal of Climate, Vol. 8,pp. 1511–1518.

Horton, E.B., C.K. Folland and D.E. Parker,2001: The changing incidence ofextremes in worldwide and centralEngland temperatures to the end of thetwentieth century. Climatic Change,Vol. 50, pp. 267–295.

Jacka, T.H and W.F. Budd, 1998: Detectionof temperature and sea ice extent

changes in the Antarctic and SouthernOcean, 1949-96. Annals of Glaciology,Vol. 27, pp. 553–559.

Janicot, S., 1992: Spatiotemporal variabilityof west African rainfall. Part II—Associated surface and airmasscharacteristics. Journal of Climate,Vol. 5, pp. 499–511.

Jauregui, E., 2003: Climatology oflandfalling hurricanes and tropicalstorms in Mexico. Atmosfera, Vol. 16,pp. 193–204.

Joint Typhoon Warning Center, 2004: JointTyphoon Warning Center best trackdataset [available online athttps://metoc.npmoc.navy.mil/jtwc/besttracks/].

Jones, P.D. and A. Moberg, 2003:Hemispheric and large-scale surface airtemperature variations: An extensiverevision and an update to 2001.Journal of Climate, Vol. 16,pp. 206–223.

Jones, P.D., K.R. Briffa, T.P. Barnett andS.F.B. Tett, 1998: High resolutionpalaeoclimatic records for the lastmillennium: integration, interpretationand comparison with generalcirculation model control runtemperatures, Holocene, Vol. 8,pp. 455–471.

Jones, P.D., M. New, D.E. Parker, S. Martinand I.G. Rigor, 1999: Surface airtemperature and its changes over thelast 150 years. Reviews of Geophysics,Vol. 37, No. 2, pp. 173-199.

Jones, P.D., T.J. Osborn, K.R. Briffa,C.K. Folland, B. Horton, L.V.Alexander, D.E. Parker andN.A. Rayner, 2001: Adjusting forsampling density in grid-box land andocean surface temperature time series.Journal of Geophysical Research,Vol. 106, pp. 3371–3380.

Kasischke, E.S., K. Bergen, R. Fennimore, F.Sotelo, G. Stephens, A. Janetos andH.H. Shugart, 1999: Contributions of1998 fires in the boreal forest toatmospheric concentrations of carbonmonoxide and methane. EOS,Transactions, American GeophysicalUnion, Vol. 80, No. 13, 141 pp.

Keeling, C.D., T.P. Whorf, M. Wahlen andJ. Vanderplicht, 1995: Interannualextremes in the rate of rise ofatmospheric carbon dioxide since 1980.Nature, Vol. 375, pp. 666–670.

Langenfelds, R.L., R.J. Francey, B.C. Pak,L.P. Steele, J. Lloyd, C.M. Trudinger and

61

References

62

C.E. Allison, 2002: Interannual growthrate variations of atmospheric CO2 andits d13C, H2, CH4, and CO between1992 and 1999 linked to biomassburning. Global BiogeochemicalCycles, Vol. 16, No. 3, 1048.

Mann, M.E., R.S.Bradley and M.K. Hughes,1999: Northern hemispheretemperatures during the pastmillennium: inferences, uncertaintiesand limitations. Geophysical ResearchLetters, Vol. 26, pp. 759–762.

Marland, G., R.A. Pielke, Sr., M. Apps,R. Avissar, R.A. Betts, K.J. Davis,P.C. Frumhoff, S.T. Jackson, L. Joyce,P. Kauppi, J. Katzenberger,K.G. MacDiken, R. Nielson, J.O. Niles,D.D.S. Niyogi, R.J. Norby, N. Pena,N. Sampson and Y. Xue, 2003: Theclimatic impacts of land surface changeand carbon management, and theimplications for climate-changemitigation policy. Climate Policy,Vol. 3, pp. 149–157.

Mears, C.A., M.C. Schabel and F.J. Wentz,2003: A reanalysis of the MSUchannel 2 tropospheric temperaturerecord. Journal of Climate, Vol. 16,pp. 3650–3664.

Met Office, 2004: The August 2003heatwave in the United Kingdom—astatement by the Met Office, Weather,Vol. 59, pp. 246.

Mo, K.C., and V.E. Kousky, 1993: Furtheranalysis of the relationship betweencirculation anomaly patterns andtropical convection. Journal ofGeophysical Research, Vol. 98,pp. 5103–5113.

Montzka, S.A., J.H. Butler, J.W. Elkins, T.M.Thompson, A.D. Clarke and L.T. Lock,1999: Present and future trends in theatmospheric burden of ozone-depletinghalogens. Nature, Vol. 398,pp. 690–694.

Montzka, S.A., J.H. Butler, B.D. Hall, D.J.Mondeel and J.W. Elkins, 2003: Adecline in tropospheric organicbromine. Geophysical Research Letters,Vol. 30, 1826.

Nakicenovic, N., O. Davidson, G. Davis,A. Grübler, T. Kram, E.L. La Rovere,B. Metz, T. Morita, W. Pepper,H. Pitcher, A. Sankovski, P. Shukla,R. Swart, R. Watson and Zhou Dadi,2000: IPCC Special Report onEmissions Scenarios. CambridgeUniversity Press, 599 pp.

National Climate Centre, China MeteorologyAgency, 2003: East Asian MonsoonMonitoring Bulletin.

Nicholls, N., 2004: The changing nature ofAustralian droughts. Climatic Change,Vol. 63, pp. 323-336.

Novelli, P.C., K.A. Masarie, P.M. Lang, B.D.Hall, R.C. Myers and J.C. Elkins, 2003:Reanalysis of tropospheric CO trends:Effects of the 1997–1998 wildfires.Journal of Geophysical Research,Vol. 108, D15, 4464.

Peterson, T.C. and R.S. Vose, 1997: Anoverview of the Global HistoricalClimatology Network temperaturedatabase. Bulletin of the AmericanMeteorological Society, Vol. 78,pp. 2837–2849.

Prather, M.J., 1996: Natural modes and timescales in atmospheric chemistry:Theory, GWPs for CH4 and CO, andrunaway growth. Geophysical ResearchLetters, Vol. 23, pp. 2597–2600.

Quayle, R.G., T.C. Peterson, A.N. Basist andC.S. Godfrey, 1999: An operationalnear-real-time global temperatureindex. Geophysical Research Letters,Vol. 26, pp. 333–335.

Ropelewski, C.F. and M.S. Halpert, 1987:Global and regional scale precipitationpatterns associated with the ElNiño/Southern Oscillation. MonthlyWeather Review, Vol. 115,pp. 1606–1626.

Ropelewski, C.F. and M.S. Halpert, 1989:Precipitation patterns associated withthe high index phase of the SouthernOscillation. Journal of Climate, Vol. 2,pp. 268–284.

Ropelewski, C.F. and M.S. Halpert, 1996:Quantifying Southern Oscillation—precipitation relationships. Journal ofClimate, Vol. 9, pp. 1043–1059.

Simpson, R.H., 1974: The hurricane disasterpotential scale. Weatherwise, Vol. 27,pp. 169–186.

Thiaw, W.M., A.B. Barnston and V. Kumar,1999: Predictions of African rainfall onthe seasonal time scale. Journal ofGeophysical Research, Vol. 104,pp. 31589–31597.

Vinnikov, K.Y., and N.C. Grody, 2003:Global warming trend of meantropospheric temperature observed bysatellites. Science, Vol. 302,pp. 269–272.

Vose, R.S., R.L. Schmoyer, P.M. Steurer,T.C. Peterson, R. Heim, T.R. Karl andJ. Eischeid, 1992: The Global Historical

Climatology Network: Long-termMonthly Temperature, Precipitation,Sea Level Pressure, and StationPressure Data. Carbon DioxideInformation Analysis Center, Oak RidgeNational Laboratory, Rep.ORNL/CDIAC-53, NDP-041, 189 pp.

Waple, A.M. and J.H. Lawrimore, 2003: Stateof the climate in 2002. Bulletin of theAmerican Meteorological Society,Vol. 84, pp. S1–S68.

Ward, M.N., 1998: Diagnosis and short-leadtime prediction of summer rainfall intropical North Africa at interannual and

multidecadal timescales. Journal ofClimate, Vol. 11, pp. 3167–3191.

World Meteorological Organization, 2002:WMO Antarctic Ozone Bulletins.[available online athttps://www.wmo.ch/web/arep/gawozobull02.html].

World Meteorological Organization, 2003:Controlled substances and other sourcegases (S.A. Montzka and P.L. Fraser).Scientific Assessment of OzoneDepletion: 2002, Global OzoneResearch and Monitoring Project ReportNo. 47, pp. 1.1-1.83.

63

References

that variability associated with ENSOinfluences regional precipitation patterns inthe tropics and mid-latitudes (Ropelewskiand Halpert, 1987).

In 2003, the year began with theequatorial Pacific Ocean in the midst of amoderate El Niño warm event (seesubchapter 4.1). This El Niño weakenedduring the late boreal winter of 2003, and

by April, SSTs in the central and easternequatorial Pacific basin had cooled to nearnormal. Despite the cessation of the ElNiño warm event and the return to near-neutral conditions in April 2003,precipitation anomalies during theremainder of 2003 continued to reflect anEl Niño–like pattern across parts of Northand South America (see subchapters 6.1and 6.3), but returned to near normal forAustralia and the South-West Pacific (seesubchapter 6.7).

Figure 2.8 shows the spatial pattern ofprecipitation anomalies determined fromthe GHCN, comparing the boreal winter of2002–2003 (DJF) annual precipitation to the1961–1990 base period. Significant wet anddry regional anomalies were present acrossthe globe during the boreal winter of2002–2003, and traditional El Niñosignatures (Ropelewski and Halpert, 1987)were evident in the observed spatial patternof these global precipitation anomalies.Wetter-than-normal regions included partsof the eastern United States and the north-west Pacific, western and central SouthAmerica, Borneo and Sumatra in Indonesia,Madagascar and the islands of the westernIndian Ocean, southern Turkey, as well asislands in the Mediterranean Sea. Negativeanomalies and drought conditions, likelyenhanced by El Niño and the associatedshift in the equatorial Walker circulation,were observed across eastern Australia, theSouth-West Pacific and Hawaii (see chapter6 for further discussion of regionalprecipitation anomalies).

Global precipitation over the oceans isoften more difficult to quantify than landareas, due to the lack of ground-truth

15

2. Global climate

Figure 2.6Global seasonal

anomalies (in °C) oflower-stratospheric

temperature (UAH andRSS). Anomalies were

determined with respectto the 1984-1990 baseperiod. The base periodwas chosen to avoid the

impacts of the volcaniceruptions of El Chichon,

Mexico, in 1982 andMt. Pinatubo,

Philippines, in 1991.

Figure 2.7Annual global

precipitation anomaliesfrom 1900 to 2003 for

land-based stations. Dataare from GHCN, and

anomalies weredetermined with respect

to the 1961-1990period.

Figure 2.8Precipitation anomalies

for the boreal winter(December to February)

of 2002-2003 fromGHCN. Positiveanomalies (wet

conditions) are shown ingreen, and negative

anomalies (dryconditions) are shown inorange. Anomalies weredetermined with respectto the 1961-1990 base

period.

A summary of the annual growth rate ofglobal climate forcing by well-mixedgreenhouse gases is shown in Figure 3.1.Although ozone (O3) and stratosphericwater vapour are additional significantgreenhouse gases, they are neither wellmixed nor well measured, and consequentlywere not included in the figure. Historicalperspectives and more detailed descriptionsof the major greenhouse gases are offeredin the following sections. The increase inthe climate forcing between 2002 and 2003,just over 0.04 W m-2, was the largest since1998, and the portion due to CO2 was thethird largest in the period of annual data(since 1958). About 90 per cent of theincrease of greenhouse gas climate forcingin the past year was caused by CO2, 5 percent by nitrous oxide (N2O), 4 per cent bymethane (CH4), and less than 1 per cent byCFCs and related trace gases, for whichincreases of some gases were balanced bydecreases of others. The increases of CO2,N2O, and CH4, and the associated climateforcing, were all larger in 2003 than in theprevious year.

3.1 Carbon dioxide (CO2)After water vapour, CO2 is the mostimportant atmospheric greenhouse gas. Forabout 10 000 years prior to the industrialrevolution, the atmospheric abundance ofCO2 was nearly constant at approximately280 ppm (ppm = parts per million by dry-air mole fraction). This abundancerepresented a balance among large seasonalfluxes (on the order of 100 Pg C year-1,where 1 Pg = 1015g) between theatmosphere and biosphere (photosynthesisand respiration), and between theatmosphere and the ocean (physicalexchange of CO2). Since the late 1800s,atmospheric CO2 has increased byapproximately 33 per cent, primarily due toemissions from combustion of fossil fuels(currently about 7 Pg C year-1) and, to alesser extent, deforestation (0–2 Pg C year-1).

Monthly mean CO2 mole fractionsfrom the NOAA/CMDL Mauna LoaObservatory in Hawaii, USA (northernhemisphere) and from the Cape Grimbaseline air pollution station in Tasmania,Australia (southern hemisphere) are plottedin Figure 3.2. Over the entire time series,the average rate of CO2 increase isapproximately 1.4 ppm per year in bothlocations. During this time, CO2 emissionsfrom fossil fuel combustion increased from2.3 Pg C in 1958 to 6.6 Pg C in 2000(Marland, et al., 2003). As a result, the CO2

growth rate has increased fromapproximately 0.8 ppm per year averagedover the first decade of measurements to1.8 ppm per year in the most recentdecade. During 2003, CO2 increased byapproximately 2.5 ppm at the Mauna LoaObservatory site (Figure 3.2a) andapproximately 2 ppm at the Cape Grimstation (Figure 3.2b), which was faster thanthe average increase over the past decade.Furthermore, the increase at the Mauna Loasite was the first time that measured CO2

growth rates had exceeded 2 ppm duringtwo consecutive years. These growth rateswere also notable given that only weak tomoderate El Niño conditions were presentin 2002 and 2003—previous rapid rises hadoccurred during far stronger ENSO events,such as 1994 and 1998.

Overall, the average increase of CO2 inthe atmosphere since 1958 corresponds toaround 55 per cent of the CO2 emitted byfossil fuel combustion (Keeling, et al.,1995), but this fraction varies from 20 percent to 90 per cent (Conway, et al., 1994;Ciais, et al., 1995). The remaining fossil fuel

18

Chapter 3

Trends in trace gases

Figure 3.1The annual growth rateof global climate forcingby well-mixedgreenhouse gases, ΔF (inW m–2 per year), overthe period 1958-2003.(Courtesy of M. Sato)

evidence that, while total global emissionsmay be changing slowly, emissions arebeing redistributed. In Figure 3.4,differences between observed polarnorthern (53°–90°N) and polar southern(53°–90°S) annual mean CH4 mole fractions(circles) are plotted as a function of time.This difference increased from 1984 to1991, dropped abruptly in 1992, and hascontinued to decrease since then. It hasbeen suggested that this abrupt change mayhave resulted from decreased emissionsfrom the former Soviet Union and easternEurope, on the order of 12 Tg CH4 peryear, that occurred after the collapse of theSoviet economy. There are now smallamounts of CH4 (2–3 Tg CH4) capturedfrom landfills and coal-mining operations indeveloped countries, however thesedecreases have been offset by increasedCH4 emissions in developing countries. Allfuture emission scenarios in the IPCCSpecial Report on Emissions Scenarios(Nakicenovic, et al., 2000) suggestincreasing rates of CH4 emission for, atleast, the next three decades.

3.3 Carbon monoxide (CO)Unlike CO2 and CH4, carbon monoxidedoes not strongly absorb terrestrial infraredradiation, but it still impacts climate throughits chemistry. The chemistry of CO affectsOH (which influences the lifetimes of CH4

and hydrofluorocarbons) and troposphericO3 (itself a greenhouse gas), so emissionsof CO can be considered equivalent toemissions of CH4 (Prather, 1996). Currentemissions of CO may contribute more toradiative forcing over decadal timescalesthan emissions of anthropogenic N2O(Daniel and Solomon, 1998).

Globally averaged, CO mole fractionsare plotted as a function of time in Figure3.5. There is a long-term decrease inglobally averaged CO, driven by a decreasein the northern hemisphere (Novelli, et al.,2003). Superimposed on the decrease was asignificant anomaly during the late 1990sdue to tropical (Langenfelds, et al., 2002)and boreal biomass burning (Bruhwiler, etal., 2000; Kasischke, et al., 2000). Becausethe lifetime of CO is relatively short (of theorder of several months), the anomalyquickly disappeared and CO returned topre-1997 levels shortly afterward. GlobalCO continued to decrease slowly, beginningin 1999, until 2002, when more large borealfires appear to have caused another

increase in CO. For the southernhemisphere, there is no long-term trend inbaseline CO measured at Cape Grim since1985, though the yearly averaged CO dataindicates a slight decrease between 2002and 2003 with values of 52.3 and 51.9nanomoles per mole, respectively.

3.4 Decreases in ozone-depleting gases in thetroposphere

Atmospheric levels of the most abundantozone-depleting gases continued to declinein the global troposphere in 2003(Figure 3.6), and the observed decreaseswere the direct result of reduced industrialproduction as mandated internationally bythe 1987 Montreal Protocol on Substancesthat Deplete the Ozone Layer. During thepast few years, ground-based measurementssuggest that none of the most abundantCFCs are currently increasing in the globaltroposphere. Although mixing ratios ofHCFC-22, a substitute for CFCs, continued toincrease steadily in 2003, those for the minor

20

Figure 3.3(a) Globally averagedmethane mole fractions(dashed line) as afunction of time. Thesolid line is thedeseasonalized trend,and symbols are annualaverages;(b) instantaneous rate ofincrease of globally-averaged CH4determined as thederivative with respect totime of the solid lineabove. Dashed lines are±1σ uncertainties;symbols are annualincrease ±1σ). (Courtesyof E.J. Dlugokencky,NOAA CMDL)

Figure 3.4Differences betweennorthern (53°-90°N) andsouthern (53°-90°S)polar annual CH4 meansplotted as a function oftime. (Courtesy ofE.J. Dlugokencky,NOAA/CMDL)

interior areas during a five- or six-month-long rainy season. During 2003, the coastalregions recorded near-normal amounts ofprecipitation. Above-average precipitationanomalies were observed in Peru andsouthern Ecuador during most of the year.Although monthly averages during Januaryand February were slightly below normal insome areas, heavy rains were blamed foroutbreaks of dengue fever and leptospirosisin Ecuador during that period. In fact,according to the International Society forInfectious Diseases, more cases of denguefever were reported during the first twomonths of 2003 in Ecuador than during allof 2002.

Precipitation surpluses were observedthroughout the year in the typically aridPatagonia region in southern Argentina. Asignificant portion of the annual anomaly inthis region came during October–December,which are normally the driest months of theyear. Southern Patagonia received 42 percent more than its climatological annualaverage of 375 mm, with 39 per cent of thatanomaly occurring during the last threemonths of the year.

North-eastern Argentina received veryheavy rainfall during April, normally theregion’s wettest part of the year. Most ofthe region’s April precipitation totalsexceeded the 90th percentile of theclimatological distribution, and some areasreceived twice their normal amount ofprecipitation. The rain sparked severeflooding that forced the evacuation of tensof thousands of people, damagedinfrastructure, caused localized outbreaks ofhepatitis and gastrointestinal diseases, andkilled as many as 25 people. Agriculture inthe province of Santa Fé was among the

worst affected sectors, because one-sixth ofits agricultural land was submerged and200 000 hectares of crops were destroyed.

Average temperatures for 2003 acrossthe continent were above normal (notshown). The largest departures wereobserved in Venezuela, north-westernArgentina, and across northern and easternBrazil. Temperatures in these areasaveraged approximately 2°C above normalfor the year. Regions with near- to slightlybelow normal temperatures included south-western Peru, northern Bolivia, and south-western Argentina. In contrast, the Peruvianhighlands experienced a cold snap in Julyas temperatures fell to -20°C during theaustral winter. Over 200 fatalities, largelyfrom pneumonia, were reported, as well aslivestock and crop losses due to theextreme cold across southern areas of Peru.

6.4 Europe6.4.1 Overview

Annual temperatures in 2003 were wellabove average across Europe, especially inwestern regions. Although cooler than 2002,annual temperature anomalies averagedover land surfaces within the area 45°–65°Nand 25°W–60°E were 0.66°C above the1961–1990 mean. There were howeversignificant regional and seasonal variations(Figure 6.9). Temperatures across theMediterranean, southern Adriatic and muchof the north-western part of the Europeancontinent, averaged over the entire year,were above the 98th percentile of the1961–1990 distribution. In general, it wasalso a dry year (Figure 6.10), especially inFebruary–March and for most of the

40

Figure 6.9European surfacetemperature in 2003expressed as percentilesof 1961-1990, using amodified two-parametergamma distribution for(a) Dec 2002-Feb 2003;(b) Mar-May; (c) June-Aug; and (d) Sep-Nov.(Sources: Jones, et al.,2001; Horton, Follandand Parker, 2001; andJones and Moberg,2003)

configuration of SST anomalies would havebeen more consistent with the oppositerainfall pattern of an anomalously dry Saheland a wet Guinea coast. However, otherconditions, such as slightly above-normalsea level pressures in the equatorial andsouthern tropical Atlantic and anomalouslystrong westerly flow at 850 hPa over theAtlantic between 5° and 12°N (Figure 6.20)

modulated the rainfall patterns and weremore typical of the wet Sahel and dryGuinea coast during 2003.

6.5.3 Eastern Africa/the Greater Horn

The 10 eastern African countries (sometimesreferred to as the Greater Horn of Africa)include Burundi, Djibouti, Eritrea, Ethiopia,Kenya, Rwanda, Somalia, Sudan, Tanzaniaand Uganda. In this region, like many otherparts of the tropics, rainfall is the mostcritical of all the climate elements due tothe heavy dependence of the economies onrain-fed agricultural activities. Rainfall in theregion can vary greatly both temporally andspatially, with a strong dependence uponthe north-south migration of the ITCZ.Extreme rainfall conditions, such as floodsand droughts, can have significant negativeimpacts on subsistence activities in thisregion.

The Greater Horn of Africa may bedivided into three distinct climate regimesbased upon rainfall characteristics. Theseare the southern sector, the equatorialsector and the northern sector. Thesouthern sector comprises of central andsouthern Tanzania and experiences arainfall maximum during December–March.The equatorial sector (Uganda, Rwanda,Burundi, Kenya, southern Somalia, northern

45

6. Regional climate

Figure 6.20July-Sep 2003

anomalous 850-hPawind speed (shading;

m s-1) and vector windsover tropical and

subtropical Africa.Anomalies are departures

from the 1971-2000base period monthly

means.

Figure 6.21Aug-Sep 2003 mean925-hPa wind speeds

(m s-1) and vector windsover tropical and

subtropical Africa.

(Figure 6.26). The Mascarene high wasnotably weaker than average overMadagascar, which resulted in a reductionin the amount of deep tropical moisturepenetrating inland from the Indian Ocean.In contrast, the above-average March 2003rains were associated with a broad low-level anomalous cyclonic circulation oversouthern Africa, which favoured theadvection of low-level Atlantic moisture intothe southern regions of the continent.

6.6 Asia6.6.1 China

During 2003, precipitation extremes were ofgreat concern in China due to their directimpact on the economy. Much of theHuaihe River valley and the Yellow Riverbasin experienced above-normalprecipitation and flooding, while southernChina was impacted by precipitation deficitsfrom summer through to autumn. Theseseasonal anomalies are reflected in theannual anomalies of precipitation(Figure 6.27).

In summer 2003, the Huaihe Rivervalley was hit by the most severe flood toaffect the region since 1991. From 21 Juneto 22 July, six major rainfall events wereobserved in the valley with precipitationtotals as high as 400 to 600 mm. Theresultant heavy flooding caused tremendouseconomic loss. The most serious damageoccurred in Anhui, Jiangsu and Henanprovinces, where up to 58 million peoplewere affected and 2 million people wereforced to be evacuated. In total, the floodscaused more than 35 billion RMB(approximately US$ 4.2 billion) of directeconomic loss.

During late August to early September,frequent heavy precipitation eventsoccurred in the mid- to lower reaches ofthe Yellow River, where precipitation totalswere again up to twice the long-termaverage. The rains resulted in severe floodsin some regions of Shanxi, Henan,Shandong, Hubei, and Sichuan provinces.During late September to mid-October,floods caused by heavy rainfall impactedupon most parts of these regions again. TheYellow River also experienced a rareautumn flood.

In contrast, southern Chinaexperienced an extended dry spell during2003, which also coincided with aprolonged heat wave. During summer, daily

temperatures above 35°C affected mostparts of southern China for more than onemonth. Maximum temperatures werebetween 38 and 40°C in the lower reachesof the Yangtze River valley. Zhejiang, Fujianand Jiangxi endured extremely hightemperatures of between 40 and 43°C. Thenumber of hot days, as well as the areaexperiencing heat waves, were both record-breaking. The heat and drought in theseregions resulted in a shortage of freshwaterand hence hydrological power. More thanfive million hectares of arable land weredestroyed, and a number of streams andrivers dried completely.

6.6.2 South-west Asia

Portions of south-west Asia, centred onAfghanistan, endured a four-year droughtduring the period 1999–2002 (Waple andLawrimore, 2003). These dry conditionspersisted into 2003, with precipitation totals

48

Figure 6.26Oct 2002-April 2003mean 850-hPa windspeeds (m s–1) andvector winds over Africaand the Indian Oceanregion.

Figure 6.27Annual precipitationpercentage anomalies for2003 in China.Anomalies are based onthe 1971-2000reference period.

The leading edge of the south-westmonsoon reached the South Andaman Seaby 16 May, prior to the normal onset date,in association with the formation of atropical storm over the Bay of Bengal. Themodified flow pattern due to the tropicalstorm caused prolonged heat waveconditions, which lasted for more than 20days over many parts of the countryduring late May and early June. Maximumtemperatures in several locations soared toabove 50°C and more than 1 500 peoplewere reported to have died due to heatstress. In Pakistan, the city of Jacobabadrecorded a maximum temperature of 52°C.The extremely hot conditions, coupledwith late onset over peninsular India,exacerbated the drought conditions in thisarea.

The first half of the monsoon seasonwas associated with slightly above-averagerainfall, while the second half wasassociated with slightly below-averagerainfall. The seasonal rainfall over India asa whole was about 105 per cent of itslong-term means. The states of Karnatakaand Kerala bore the brunt of the maximumrainfall deficiency (Figure 6.32). Theseregions had also received below averagerainfall during the previous three to fouryears, and hence severe water shortageswere reported not only in these areas, butalso in the downstream regions due toreduced river levels.

The summer monsoon began itswithdrawal over Pakistan and westernRajasthan around 17 September,approximately two weeks later than thenormal date. However, the withdrawalprogressed briskly and hence most otherparts of the country experienced normalmonsoon departure dates, before itwithdrew from the entire country by 15October.

6.6.5 East Asian monsoon

The South-East Asian summer monsoon in2003 was weaker than normal. Themonsoon onset started over the SouthSouth China Sea during the fifth pentad ofMay and covered all of the South China Seaby the first pentad of June (Figure 6.33).

The leading edge of the summermonsoon shifted to the Yangtze and theHuaihe River basins in the fifth pentad ofJune and persisted for about one month(Figure 6.34), bringing heavy precipitationto this region. In late July, the leading

51

6. Regional climate

Figure 6.32Spatial patterns of theseasonal mean rainfallanomalies (% of long-term mean) over India

during the monsoonseason of 2003.

Figure 6.33Variation of pentad-

averaged zonal andmeridional wind (in

m s 1) and the valuebetween the potential

pseudo-equivalenttemperature and 335Kat the 850hPa level (inK) over the area 10°N-20°N, 110°E-120°E,from the first pentad of

April 2003 to thesecond pentad of

October 2003.

RAINFALL

The Australian El Niño-related droughtlasted from March 2002 through January2003 (Bureau of Meteorology, 2003). Duringthis 11-month period, 90 per cent of thecountry received below median rainfall,with 56 per cent of the country receivingrainfall in the lowest 10 per cent ofrecorded totals (i.e., decile 1). This was thesixth greatest extent of decile-1 rainfall forany 11-month period on record (Australia-wide rainfall records commenced in 1900).

However the remainder of 2003 saw ageneral return to near-normal rainfallconditions (Figure 6.35), resulting in themean Australia-wide rainfall total (476 mm;Figure 6.36) being 4 mm above the1961–1990 normal (472 mm) and 38 mmabove the long-term median (438 mm).Overall, 61 per cent of the countryexperienced rainfall that was above themedian, compared to only 18 per cent in2002. Similarly, only 1.2 per cent of thecountry experienced decile 1 rainfall during2003, compared to 36 per cent the yearbefore.

Although the meteorological droughthad largely subsided by mid-2003,hydrological drought remained. Low waterstorages led to the establishment of waterrestrictions for Australia’s two largest cities(Sydney and Melbourne), and the nationscapital, Canberra. Ongoing dry conditionsin south-eastern Queensland saw storagedam levels fall 20 to 30 per cent ofcapacity, with some less than 10 per cent.Irrigation allowances were reduced orsuspended in many major catchments in

eastern Australia, impacting heavily onagricultural activities.

TEMPERATURES

The severity of, and slow recovery from,the 2002-2003 drought may have beenimpacted by the higher-than-normaldaytime and nighttime temperatures(Nicholls, 2004). The Australia-widemaximum temperature anomaly for 2003was the sixth highest on record (recordscommenced in 1910), with a value of0.65°C above the 1961–1990 mean (Figure6.36). Regional anomalies were as large as+2.5°C (eastern Queensland) (Figure 6.37).The Australia-wide minimum temperatureanomaly of +0.59°C was the fourth higheston record (Figure 6.36), and hence theAustralia-wide mean temperature anomalyof +0.62°C was the sixth highest on record.The significantly above-normal Australiantemperatures of 2003 were the result of afairly typical post-El Niño warm period, asheat from the November–December 2002peak of the event dissipated over thefollowing months, overlaid on the longer-term warming trend.

Australia’s high temperatures wereexemplified by a widespread heat waveduring the austral spring. September recordswere broken in New South Wales, SouthAustralia, Victoria and Western Australia,with very high temperatures in the NorthernTerritory and Queensland. In New SouthWales alone, nine locations in the north andwest exceeded the previous Septemberrecord of 38.8°C. Another heat wave innorth-west Western Australia broke the all-time Australian September record (42.8°C),when West Roebuck recorded 43.1°C on 27September.

The hot and dry conditions leadinginto early 2003 led to major bushfires inNew South Wales, Victoria, the AustralianCapital Territory, Western Australia andTasmania. The largest occurred whenlightning strikes started fires in the alpineareas of New South Wales, the AustralianCapital Territory and Victoria which laterjoined together to form a massive complexthat burned for 59 days during January andFebruary. This fire was the third largest firein south-eastern Australian history, after the“Black Friday” fires of January 1939, andthe “Black Thursday” fires of February 1851.Four people perished when the firesentered the suburbs of Canberra.

53

6. Regional climate

Figure 6.37Maximum temperature

anomalies acrossAustralia during 2003,

with respect to the1961-1990 base

period.

renewable energy – human health – food security – reducing poverty – disaster preventionrenewable energy – human health – food security – reducing poverty – disaster prevention

World Meteorological Organization

For information about WMO, please contact:Communications and Public Affairs Office

World Meteorological Organization7bis, avenue de la Paix - P.O. Box 2300 - CH 1211 Geneva 2 - Switzerland

Tel.: (+41-22) 730 83 14 - 730 83 15 - Fax: (+41-22) 730 80 27

E-mail: [email protected] - Website: www.wmo.int

For information about the Global Climate System Review, please contact:World Climate Programme Department

World Meteorological Organization7bis, avenue de la Paix - P.O. Box 2300 - CH 1211 Geneva 2 - Switzerland

Tel.: (+41-22) 730 82 68 - Fax: (+41-22) 730 80 42

E-mail: [email protected] - Website: www.wmo.int/web/wcp/wcp_prog.htm

the global climate system review 2003 - vsamp · of the global climate system review provides a...

Documents