NEWS NEWSVol. 15, No. 2 May 2005Global Energy and Water Cycle Experiment

World Climate Research Programme(A Programme of WMO, ICSU and IOC)

GEWEX AND EXTREME EVENTSObservations

Assimilation and Reanalysis

Extreme Events

Planning and Response

Newsletter ContentsCommentary: A More Focused Agenda for 2

GEWEX Application Efforts

Recent News of Relevance to GEWEX 3

5th Int'l Scientific Conference on the Global 4 Energy and Water Cycle Abbreviated Program

A Focus on Extremes in GEWEX 5

Use of Indices in Characterizing Precipitation 5 Extremes: A European Example

GEWEX SSG Chair Receives Highest 6 NASA Honor

Precipitation Extremes over Viet Nam 7

A Water Balance Investigation of the 2002 8 Drought in the Murray-Darling Basin

European Precipitation Extremes: Can We 10 Expect More?

U.S. Historic Drought Characteristics 12 Estimated Using N-LDAS

GEWEX Relevant Publications of Interest 14Workshop/Meeting Summaries:– IGWCO/GEWEX/UNESCO Workshop on 15

Trends in the Global Water Cycle– Fourth CEOP Planning Meeting and 16

First IGWCO Workshop– Twenty-Sixth JSC Meeting 17– Workshop on the Applicability of Climate 18

Research for Water Resource Managementin Semi-Arid and Arid Regions

GEWEX/WCRP Meetings Calendar 19

5th International Scientific Conference on the Global Energy and Water Cycle

160 Oral Presentations and 175 Posters on the Program (see page 4)

May 20052

COMMENTARY

A MORE FOCUSED AGENDAFOR GEWEX APPLICATION EFFORTS

Soroosh Sorooshian, ChairmanGEWEX Scientific Steering Group

When GEWEX was originally conceived by theWorld Climate Research Programme (WCRP) inthe late 1980s, it established a number of scientific,observation, and modelling objectives to guide itscontributions to society. With respect to the latter,the stated objective has been to "develop theability to predict the variations of global andregional hydrological processes and water re-sources, and their response to environmentalchange." During the early years of GEWEX, andfor obvious reasons of the required progressionfrom science to applications, other GEWEX ob-jectives received higher priority. The ContinentalScale Experiments (CSE), under the GEWEX Hy-drometeorological Panel (GHP) were given the lead,but their results were mixed and on the average didnot meet expectations.

To better address this deficiency, GHP estab-lished the Water Resources Application Project(WRAP) with the mandate to explore how best thescience, knowledge, and information generated byGEWEX can be transferred to the water resourcesmanagement community. WRAP, under the ableleadership of Lawrence Martz, has done a com-mendable job in sponsoring a number of sessionsat various national and international conferences,and sharing the results of the various applicationprojects. The experience thus far has shown that:

• While a number of water resources casestudies and application projects have beensupported [especially by the GEWEX Con-tinental-scale International Project (GCIP)/GEWEX Americas Prediction Project (GAPP)],it is not clear to what extent these projectsbenefited directly from GEWEX products(i.e., science, modelling and observations).Would the results and outcomes of theseprojects have been any different withoutGEWEX?

• While "politically appealing," the originalGEWEX objective was simply too broadand provided no realistic roadmap. Further-more, prior to WRAP, there was little inputfrom the user side to articulate a realisticplan.

• The resources required to carry out thisobjective to a minimum level have neverbeen sufficient. Allocating a few hundredthousand dollars or euros out of the sciencebudget for GEWEX-related research will notdo the job except to give a "good impression."

• Finally, GEWEX alone cannot fulfill this ob-jective, even when it is defined more narrowly.

In GEWEX Phase II we are attempting to ad-dress the above issues. We have revised this objectiveby targeting the operational hydrometeorologicalservices and hydrologic research programs as theprimary user communities. We believe that it ismuch more realistic to infuse GEWEX productsinto hydrologic models and hence to improve thequality and reliability of hydrologic predictions socritical to the water resources management commu-nity. For example, Global Land Data AssimilationSystem and Global Soil Wetness Project productsshould be ideal candidates for initializing and up-dating the states of the hydrologic rainfall runoffmodels, as well as filling the void in some of theinput data requirements of distributed hydrologicmodels. The high resolution land-surface models,a GEWEX achievement, can serve as a bridge tolink longer term (up to seasonal-to-interannual) cli-mate predictions to ensemble hydrologic forecaststo guide reservoir operators and drought manage-ment.

This new objective poses a challenge for WRAPand GEWEX/WCRP. WRAP must help the CSEsteam up with the appropriate hydrologic servicesand develop some demonstration and pilot projectswithin their geographic regions. This is alreadyhappening through a partnership between GAPPand the U.S. National Weather Service Office ofHydrologic Development.

Perhaps a bigger challenge is to go beyondthese demonstration projects and make a differ-ence at the global scale. GEWEX needs closercollaboration with those hydrologic and waterresources programs that coordinate hydrologicservices and information gathering and dissemi-nation among member nations. The WorldMeteorological Organization's (WMO) Hydrologyand Water Resources Division and the UnitedNations Educational Scientific and CulturalOrganization's (UNESCO) International Hydrol-ogy Programme are the two obvious partners forGEWEX.

3May 2005

ISLSCP Initiative II WorkshopOn 3–5 May 2005 an international workshop

was held at the University of Maryland, BaltimoreCounty, to examine the results of early analysiscarried out with the new International Satellite LandSurface Climatology Project (ISLSCP) Initiative II

GEWEX Outreach

In April, materials on GEWEX, the Coordi-nated Enhanced Observing Period, the Global WaterSystem Project, and WCRP were distributed viainformation tables organized by the Integrated Glo-bal Observing Strategy-Partners and the Committeeon Earth Observation Satellites at the Commissionfor Sustainable Development Meeting in New YorkCity, New York, U.S.A. and the European Geo-physical Union Meeting in Vienna, Austria.

iPILPS First International Workshop

The isotopes in the Project for Intercomparisonof Land-surface Parameterization Schemes (iPILPS),which is evaluating and refining isotope-enabled land-surface schemes in weather, climate and Earth systemmodels was “kicked off” at a highly successful 5-dayinternational workshop hosted at the Australian NuclearScience and Technology Organization in Sydney,Australia in April 2005. This workshop featuredpresentations from respected international and localresearch leaders, and discussions aimed at evaluatingand improving representations of stable water isotopeprocesses in the land surface schemes of globalclimate models. Further information can be found onthe iPILPS web page: http://ipilps.ansto.gov.au/.

As a first step, WRAP organized a workshopin Cairo between GEWEX/WCRP and UNESCO’sWater and Development Information for Arid Lands-A Global Network (G-WADI) Initiative to assessthe hydroclimate prediction needs of the world’ssemi-arid regions (see page 18). Furthermore, therecent hydrologic modelling workshop in April heldin Roorkee, India assessed the hydrologic model-ling needs of semi-arid regions and the barriers tousing more advanced models, especially in devel-oping countries.

While we are making good headway withUNESCO, our engagement with WMO’s Hydrol-ogy Programme has yet to develop. There is agrowing need to show the benefits of mutual co-operation between the two WMO programmes sothat they can better serve the hydrologic predictionneeds at the global scale. In addition to the WRAPefforts, the Directors of WCRP and the WMOHydrology and Water Resources Programs need toinitiate a dialogue.

GEWEX Reaches Out to the EuropeanScience Community

Thanks to the efforts of Dr. Peter van Oevelen,a special side session on GEWEX was organizedat the European Geophysical Union in Vienna inApril 2005. The session featured presentations onmany of the major projects within GEWEX aswell as an overview of the Integrated Land Ecosys-tem-Atmosphere Processes Study (iLEAPS). It ishoped that these efforts will encourage more mem-bers of the European scientific community to engageactively in addressing science questions that arecentral to the GEWEX endeavor. In addition, Ms.Catherine Michaut from the WCRP/CoordinatedObservation and Prediction of the Earth System(COPES) Project Office distributed material onWCRP and its projects to attendees. A planningsession for the GEWEX Hydrometeorology PanelTransferability study was organized by Dr. EugeneTackle and chaired by Dr. William Gutowski.

NOTICE

Under "GEWEX AMS Honorees" in the Feb-ruary 2005 issue of GEWEX News, John Roads,Chair of the GEWEX Hydrometeorology Panel,should also have been recognized as a new AmericanMeteorological Society Fellow.

RECENT NEWS OF RELEVANCETO GEWEX

data sets, and to discuss plans for the future ofISLSCP within GEWEX. The meeting, which drewscientists from the U.S. and Europe, endorsed theusefulness of the Initiative II products for modeldevelopment and validation, and identified someareas where improvements and metadata are needed.

Discussions about ISLSCP Next explored thepossibility of developing a data set that was tar-geted to a specific scientific thrust and ways toentrain European and other international activitiessuch as the Global Monitoring for the Environmentand Security into the process.

May 20054

5th INTERNATIONAL SCIENTIFIC CONFERENCE ON THE GLOBAL ENERGY AND WATER CYCLE

Costa Mesa, California, USA

ABBREVIATED PROGRAM(for the complete program, see http://www.gewex.org/5thconf.htm)

Conference Themes and Presentation Foci:Theme 1: Clouds and their Effect on the Radiation

BudgetTheme 2: Predictions for Water ManagementTheme 3: Roles of Land Fluxes in Water and

Energy CyclesTheme 4: The Role of Modelling in Predictability

and Prediction StudiesTheme 5: New Strategies for Characterizing and

Predicting Energy and Water BudgetsTheme 6: Measuring and Predicting PrecipitationGAPP: GEWEX Americas Prediction ProjectCEOP: Coordinated Enhanced Observing Period

Sunday, 19 June 20051700 – 2000: Early Registration and Icebreaker

Monday, 20 June 20050700 – 0845: Conference Registration0845 – 1025: Plenary Session0845 – 0900: Introductions0905 – 1025: Key Note Speakers

0905 – 0925: COPES: An Integrative Strategy forWCRP, Peter Lemke (Alfred WegenerInstitute, Germany)

0925 – 0945: Priorities for US Water Cycle Research(TBC), Kathie Olsen (Office of Scienceand Technology Policy, USA)

0945 – 1005: Overview of NASA’s Satellite Observa-tions of Water in the Earth System,Jack Kaye (NASA Science MissionDirectorate, USA)

1005 – 1025: The Contributions of ESA to Energyand Water Cycle Observations andResearch, Einar-Arne Herland (EuropeanSpace Agency/ESTEC, The Netherlands)

1025 – 1040: Morning Break1040 – 1220: Simultaneous Sessions, Themes 1 and 21220 – 1400: Lunch1400 – 1530: Simultaneous Sessions, Themes 1 and 21530 – 1600: Afternoon Break1600 – 1700: Simultaneous Sessions, Themes 1 and 21700 – 1800: Poster Viewing, Themes 1 and 2

Tuesday, 21 June 20050830 – 1020: Plenary Session1020 – 1040: Morning Break1040 – 1220: Simultaneous Sessions, Themes 1 and 31220 – 1400: Lunch1400 – 1530: Simultaneous Sessions, Themes 2 and 31530 – 1600: Afternoon Break1600 – 1700: Simultaneous Sessions, Themes 3 and 41700 – 2000: Poster Viewing, Themes 3, 4 and GAPP1800 – 2000: GAPP Townhall Meeting and Reception

Wednesday, 22 June 20050830 – 1010: Plenary Session1020 – 1040: Morning Break1040 – 1220: Simultaneous Sessions, Themes 3 and 41220 – 1400: Lunch1400 – 1530: Simultaneous Sessions, Themes 3 and 41530 – 1800: Side Sessions for Working Groups1900 – 2130: Banquet

Speaker: Wes Bannister, Chairman,Board of Directors, Metropolitan WaterDistrict of Southern California, Topic:Planning for Uncertainty in No UncertainTerms: A Water Agency’s Mission toEnsure Reliability

Thursday, 23 June 20050830 – 1020: Plenary Session1020 – 1040: Morning Break1040 – 1220: Simultaneous Sessions, Themes 4 and 51220 – 1400: Lunch1400 – 1530: Simultaneous Sessions, Themes 5 and 61530 – 1600: Afternoon Break1600 – 1700: Simultaneous Sessions, Themes 5 and 61700 – 2000: Poster Viewing, Themes 5, 6 and CEOP1800 – 2000: CEOP Townhall Meeting and Reception

Friday, 24 June 20050830 – 1010: Plenary Session1020 – 1040: Morning Break1040 – 1250: Simultaneous Sessions, Themes 5 and 61240 – 1400: Lunch1400 – 1530: Panel Discussion: GEWEX PHASE II:

Providing Leadership for COPES,WCRP,IGBP and the Climate and EarthObservations Communities

1530 – 1540: Conference Close

5May 2005

A FOCUS ON EXTREMES IN GEWEX

Ronald StewartMcGill University, Canada

Importance of Extremes: One of the most criticalaspects of the water and energy cycle is the occur-rence of extremes such as droughts and extended wetperiods. Not only do they have enormous impactswhen and where they occur but they are also funda-mental features of the climate system.

Extremes and GEWEX: The strategy of the GEWEXHydrometeorology Panel (GHP) has concentrated onassembling relevant data sets, addressing long-termwater and energy budgets, assessing sources andsinks of moisture, and interacting with the waterresource community. GHP, along with the otherGEWEX components, is now ready to examine ex-tremes in a systematic manner. The primary objectivesof this effort will be to better understand and modelthe occurrence, evolution, and role of extremes withinthe climate system and to contribute to their betterprediction. These objectives will be addressed bydetermining the extent to which similar processes areresponsible for extremes in different regions, andunderstanding the processes that link extremes indifferent regions and how they may be changing.

In part these activities will consider: (1) extremesas a natural follow-on to current water and energybudget related studies; (2) the Coordinated EnhancedObserving Period (CEOP) Phase I data period as acase study; and (3) trends in the occurrence ofextremes through the analysis of long-term records.

To some extent a study of extremes would buildon existing initiatives. In particular, it could expandthe Water and Energy Budget Study (WEBS) with afocus on droughts and extended wet periods ratherthan on the average conditions occurring in thoseregions. During Phase I of CEOP (2001–2004),numerous extremes occurred, including a long-lastingdrought over western North America, a devastatingheat wave in Europe, and periods of drought as wellas flooding over Asia, South America, Australia, andAfrica. On longer time scales, multi-decade recordsassembled by GEWEX and other groups, includingre-analysis efforts, can all be exploited to focus ontrends in extremes.

Extremes are important to other climate and di-saster related initiatives. Linkages will be developedwith other WCRP projects and international effortsincluding disaster research.

USE OF INDICES IN CHARACTERIZINGPRECIPITATION EXTREMES:

A EUROPEAN EXAMPLE

Olga Zolina1, Clemens Simmer1

Alice Kapala1, and Sergey Gulev2

1Meteorologisches Institut, Universitaet Bonn, Germany, 2P. P. Shirshov Institute

of Oceanology, Moscow, Russia

Accurate estimates of the changes in extremeprecipitation from station measurements are veryimportant for validation of model experiments aimedto quantify trends in extreme precipitation underglobal warming. The linear trend estimates in ex-treme precipitation indices over Europe showincreasing probability of the occurrence of heavyprecipitation for both stations and models. Al-though these indices may be qualitatively consistentwith each other, they may, however, show somedisagreement in their quantitative estimates of lineartrends (Frei and Schar, 2001; Groismann et al.,2005; Klein Tank and Koennen, 2003). Moreover,when missing values of daily precipitation areinhomogeneously distributed they contribute addi-tional uncertainty in the trend estimates.

In this study, we used daily rainfall data fromthe Royal Netherlands Meteorological Institute Eu-ropean Climate Assessment data set along with thecollections of the Russian National Climate DataCenter and the German Weather Service (Zolina etal., 2004). These data consist of 295 stations from1804–2003 with 96 station records longer than 100years and characterized by the homogeneity of theobservational practices. Using 22 stations with recordswithout gaps we derived the distributions of thenumber of missing values and of the durations ofgaps, and developed sampling models which wereused for the homogenization of centennial dailyrecords for all 96 stations with complete records.From the homogenized time series we derived dif-ferent precipitation indices and further analyzedlinear centennial trends in these indices. The threeindices used were the exceedance of 95% thresh-old (G95) (Groismann et al., 2005), the percentageof seasonal total during very wet (>95%) days(R95) (Klein Tank and Koenen, 2003) and the 95%percentile (p95) of precipitation from the estimatedGamma distribution for the daily precipitation (Zolinaet al., 2004).

Trend estimates (in percentage per decade) forR95 may exhibit locally significant differences with

May 20056

those for p95, while trends in G95 are very closeto p95. During winter, trends in p95 and R95 areprimarily positive in Central and Eastern Europe(2–7% and 3–10% per decade, respectively). In 43locations the indices demonstrated significant changes.However, the trends in all three indices are signifi-cant and show the same sign at only 19 locations.In Western Europe in summer all three indicesshow negative trends in the northern part and posi-tive changes in the Alps. Summer trends of theseindices in Eastern Europe are less consistent. Sig-nificant linear trends with the same sign exist onlyfor 12 of 34 locations where either index demon-strated significant changes. Considerable disagreementis also observed in Northern Europe where R95indicates stronger negative changes.

In the figure above the estimates of linear trendsin annual values of p95 are shown for the 22 loca-tions where all three indices (p95, R95, G95) computedfrom the homogenized time series are significantaccording to different statistical criteria, such asthe Student t-test, the Hayashi criterion and theWilcoxon test, and are simultaneously unaffectedby the sampling density. In other words, these mapsshow the locations where trends in heavy Europeanprecipitation can be accepted with confidence. These

Estimates of linear trends in annual values of (A) p95and (B) p99 in the locations where all three indices(p95, R95, G95) computed from the homogenized timeseries are significant according to the chosen criteria.

trends are positive in most of Eastern Europe,where the strongest changes range from 3 to 5%per decade, and are somewhat weaker in CentralWestern Europe. The trends in p99 (see Panel B ofthe figure) are significant in 12 locations only, withprimarily positive tendency and maxima of 3 to 4%change per decade in Northern European Russia. Ageneral conclusion can be made that the use ofmore objective indices based on estimated PDFsfor daily precipitation, appears to be the best strat-egy for estimating long-term tendencies in extremeprecipitation.

Acknowledgements. This study was supported bythe North Rhine-Westphalian Academy of Scienceunder the project “Large Scale Climate Changesand their Environmental Relevance,” RFBR (Grants03-05-64506 and 03-05-20016) and the NATO-SfP-981044 project “Extreme precipitation events: theirorigins, predictability and impacts.”

References

Frei, C., and C. Schär, 2001. Detection probability of trends inrare events: theory and application to heavy precipitation in theAlpine region. J. Climate, 14, 1568–1584.

Groisman, P. Y., R. W. Knight, D. R. Easterling, T. R. Karl,G. C. Hegerl, and V. N. Razuvaev, 2005. Trends in intenseprecipitation in the climate record. J. Climate, 18, in print.

Klein Tank, A. M. G., and G. P. Koennen, 2003. Trends inindices of daily temperature and precipitation extremes in Eu-rope, 1946-99. J. Climate, 16, 3665–3680.

Zolina, O., A. Kapala, C. Simmer, and S. K. Gulev, 2004.Analysis of extreme precipitation over Europe from differentreanalyses: a comparative assessment. Glob. Planet. Change,44, 129–161.

GEWEX SSG Chair ReceivesHighest NASA Honor

On 27 April 2005, Prof. SorooshSorooshian, chair of the GEWEXScientific Steering Group, receivedthe Distinguished Public ServiceMedal, the highest honor that theNational Aeronautics and SpaceAdministration (NASA) awards tosomeone who is not a governmentemployee. The award is granted only

to individuals whose distinguished accomplishmentscontributed substantially to the NASA mission. Prof.Sorooshian received this award for his distinguishedrecord in providing scientific leadership for globalwater cycle research and assuring that NASA scienceis well integrated into international programs.

7May 2005

PRECIPITATION EXTREMESOVER VIET NAM

Dang Thi Mai and Nguyen Tan Thanh

Mid-Central Regional Hydro-Meteorological Center of Viet Nam

Historically, the mid-central region of Viet Namexperiences annual severe weather events, includ-ing typhoons, hurricanes, floods, and droughts thatlead to the loss of life and property. In recentyears, these events have become more severe andare occurring in new areas. This article describesextreme rain events that occurred on 1–6 Novem-ber 1999 and 23–27 November 2004, where rainrates exceeded historical expectations and forecastinformation, leading to unusually large impacts

The mid-central region of Viet Nam is locatedfrom 14o32' to 18o06'N and 105o37' to 109o05’E,and includes five provinces, Quang Binh, QuangTri, Thua Thien Hue, Quang Nam, Quang Ngaiand Da Nang City. This coastal region contains anarrow flat area with some low hills on the eastand a part of the Truong Son high mountain ridgeon the west that is cut by many mountain passes.The topography is quite steep from west to eastand the rivers are short and flow rapidly. Mon-soons and Truong Son Mountain Range, playimportant roles in the precipitation regime. Thenortheast monsoon, the southwest monsoon andthe prevailing trade winds give the region twoseasons: dry (January to August), and rainy (Sep-tember to December). Rainfall is heavier in themountain areas than in the flat areas.

Event 1: 1–6 November 1999. On 1 Novem-ber a northeast monsoon and a cold front occurredat 18–19oN and a tropical convergent zone oc-curred in the south at 11–12oN. In the upperatmosphere the easterly wind field developed strongwinds with velocities of over 10 m/s. The humid-ity was high for most of the mid-central region.All stations from Thua Thien Hue to Quang Ngairecorded daily rainfall averages more than 100 mmfor 1–6 November. For 1–3 November, more than50% of the stations recorded rainfall greater than300 mm (about 20% more rainfall than 500 mm fora 24-hour period). In particular, Hue City expe-rienced 1,422 mm of rainfall from 6 a.m. on 2November to 6 p.m. on 3 November with apeak rainfall intensity of 120 mm every 60minutes. On 2 November Nam Dong had a maxi-mum daily rainfall of 593 mm, almost 20% morethan the previous maximum daily rainfall record

(500 mm). Over 700 deaths and a loss of propertytotaling $250 million resulted from this event.

Event 2: 23–27 November 2004. For 23–25November, the weather in the mid-central regionwas influenced in the southern part by high pres-sure and the typhoon Muifa in the northern part.From 25–27 November a strong northeast mon-soon event impacted the area and in the upperlevels the easterly winds were very strong at 10–15 m/s. Three hundred people narrowly escapeda flash flood which occurred in the Suoi Luong, asmall stream in the suburb of Da Nang City. Dur-ing the same period a flash flood in Tay Tra(Quang Ngai province) caused a landslide that killedfive people. A total of 37 people were killed in themid-central region by flooding during this 23–27November event.

The analysis of these severe rainfall eventsshowed that the interaction between the northeastmonsoon and typhoon circulation (Intertropical Con-vergence Zone) created unstable air masses in theSouth China Sea. Easterly winds carried that un-stable air mass with high amounts of precipitablewater towards land, where it was then blocked bythe Truong Son Range, resulting in heavy rainfall inthe region’s mid-central territory.

Analysis also shows that in recent years, flood-ing and in particular, flash flooding, is occurringmore frequently and appears to be related to defor-estation. Heavy rainfalls can be predicted 12–24hours in advance and flooding 6–12 hours in ad-vance. However, Viet Nam does not have techniquesavailable to predict flash floods. The informationavailable to forecasters in Viet Nam includes syn-optic maps of the surface to 500 hPa height,meteorological data, satellite images, radar prod-ucts and numerical models. The precipitationforecasted in numerical models averages 50–100mm less than actual precipitation. Additional raingauges are needed in the mid-central region surfaceobservation network, especially in the narrow riverbasins where flash flooding often occurs, as wellas more satellite rainfall measurements.

Heavy rain is a very complex phenomenon withevery event having different characteristics, whichrequire detailed analysis. To effectively understand,monitor and predict these extreme events, the Na-tional Hydro–Meteorological Service of Viet Namand the Mid-Central Regional Hydro–Meteorologi-cal Center (like many other meteorological servicesin other parts of the world) are seeking to improvetheir capacity and develop a long-term moderniza-tion plan for rainfall monitoring and prediction.

May 20058

A WATER BALANCE INVESTIGATIONOF THE 2002 DROUGHT IN THE

MURRAY-DARLING BASIN

C. S. Draper, and G. A. Mills

Bureau of MeteorologyMelbourne, Australia

In 2002, southeast Australia experienced one ofits worst droughts on record, resulting in the wide-spread loss of rural livelihood as well as the lossof several lives. One area that was particularly hardhit by the drought was the Murray-Darling Basin(MDB). The MDB is one of Australia’s majoragricultural regions, providing 40% of Australia’sagricultural production from just 14% of its landmass. The Basin is largely semi-arid, and has lim-ited available water supplies, with over 80% ofdivertible surface water being consumed within theBasin. The climate of the Basin is typified by longdry spells, punctuated by episodic wet events. Theregular occurrence of long dry spells places enor-mous pressure on the Basin’s water resources. Asa result, both the MDB and the Australian economyare sensitive to climatic fluctuations in the region.

The atmospheric water balance of the MDB iscurrently being investigated as a part of the GEWEXMDB Continental-Scale Experiment. The investiga-tion is based on output from the Australian Bureauof Meteorology’s operational forecast model, theLimited Area Prediction System (LAPS). The 2002drought has received particular attention in thisinvestigation, with the goal of determining the im-portant processes occurring in the water cycle, aswell as testing the performance of LAPS duringextreme dry events. What follows is a summary ofthe 2002 drought and its impact on southeast Aus-tralia, and an overview of the methodology of theMDB Water Balance Project and its main results.One of the most interesting findings of the Projectis that LAPS is not accurately representing extremedry events, and the reasons for this—and likelyconsequences—have also been presented.

The 2002 drought was primarily the result ofstrong El Niño conditions in the later half of theyear, which suppressed precipitation across south-east Australia. However, the low precipitation (inthe lowest decile) does not alone account for theseverity of the drought. In fact annual precipitationwas lower several times in the last half century(e.g., 1982 and 1994), during which drought symp-toms were less severe than in 2002. The impactsof low precipitation in 2002 were greatly enhanced

by two factors. Firstly, the preceding few yearshad been consistently dry, and secondly, recordhigh temperatures were experienced throughout 2002.The all-Australia maximum temperature anomaly was+1.65°C for the 9 months from March to Novem-ber, which is the warmest on record by a marginof +0.63°C (Watkins, 2002).

The 2002 drought had a major impact onAustralia’s agricultural sector, which was felt bythe larger economy. The Australian Bureau ofAgriculture and Resource Economics estimated thatagricultural production in 2002 was 57% of thetonage produced the previous year. Production inthe state of New South Wales, 75% of which iswithin the Murray-Darling Basin, was estimated tohave dropped from 8.9 million tons in 2001 to 1.3million tons in 2002 (Watkins, 2002). The Gover-nor of the Reserve Bank of Australia estimated thatthe dry conditions reduced Australia’s annual GrossDomestic Product (GDP) by one percentage point(Watkins, 2002).

The 2002/03 bushfire season was particularlysevere. In December there were more than 58individual fires burning in New South Wales, whichwere estimated to have caused $100 million ofdamage (Watkins, 2002). In early 2003 the situa-tion worsened and 3 million hectares were burnt insoutheast Australia during January and February,resulting in huge stock and infrastructure losses.Fires in and around Canberra in mid-January burned500 houses and resulted in five deaths (Council ofAustralian Governments, 2004). The Canberra fireswere the most expensive ever in terms of insurancepayments.

The results of operational forecast models areof particular interest during extreme events, suchas the drought of 2002. The MDB Water BalanceProject is based on LAPS model output at a 3-hour temporal resolution, with a horizontal resolutionof 0.375°. There are two forecasts made each day,at base times 12 and 00 UTC. For each time step,the relevant fields have been calculated from theaverage of the most recent forecast from each ofthe base times. The average daily value per monthof each water balance term is shown in the figureon page 9 for the year preceding the end of the2002/2003 summer.

A number of interesting observations can bemade in the figure on page 9. The low precipitationin 2002 can be observed by comparison to otheryears (1999–2004 have been calculated, but are notshown). In February 2003 there is a rapid increasein precipitation, indicating the end of severe drought

9May 2005

conditions, due partly to the weakening of El Niñoconditions. This follows a recent trend towardshigh February rainfall in Australia. February is theonly month in the figure above during which pre-cipitation exceeded evaporation. An evaporationexcess (over precipitation) is unusual during thecooler months and all other years studied haveshown a consistent, although not necessarily large,precipitation excess in winter. While evaporationexcesses are typical during the warmer months, themonthly excesses calculated for 2002 are muchhigher than in other years, with January being par-ticularly extreme. The divergence is not a leadingterm in the water balance, indicating that recyclingof evaporated water is providing source vapor forprecipitation, rather than convergence of water fromexternal sources.

Perhaps the most striking feature in the figureis the large negative residual calculated during springand summer. These residuals were not unexpected,as it is well known that budget studies based onnumerical weather prediction (NWP) models, andassociated analyses contain an inherent error whichis not small (Kanamitsu and Saha, 1996). Thiserror is related to the systematic tendency of themodel to drift towards a ‘model climate,’ whichis determined by its internal physics andparameterizations. Water balance studies based onseveral NWP models (and associated analyses) aroundthe globe have calculated significant residuals (e.g.,Roads et al., 2002). The negative residual hereindicates that LAPS has a tendency to accumulateexcess moisture in its atmosphere, which must thenbe artificially removed (nudged) during each as-similation cycle. This process is clearly illustrated

in the diurnal cycle of the forecast atmosphericwater storage, which shows a general increasingtrend through the 24 hours of each forecast, andthen drops abruptly to meet the newly assimilatedvalue at the introduction of the next forecast. Theresiduals during spring and summer are large, bothin comparison to the other water balance terms,and to the monthly residuals calculated for othertime periods. The residual is one of the leadingterms across the months in question, and the re-sidual term for January 2003 contributes half of thenet residual for 2003.

Preliminary investigations indicate that the ten-dency in LAPS towards accumulating excessatmospheric moisture during extreme dry events iscaused by an over-prediction of evaporation inthese conditions, and the reasons for this are cur-rently being investigated. Possible contributors arethe atmospheric moisture assimilation process, themethod used to update soil moisture during eachassimilation cycle, and/or the land surface param-eterization of evaporation. A positive evaporationbias during dry events could have significant con-sequences for the performance of LAPS, althoughthese consequences are difficult to predict. It wouldbe reasonable to expect a cool bias to occur,which would have many consequences. The excessevaporation would certainly affect the humidity field,which would impact the skill of fire weather pre-dictions, as these become extremely sensitive tosmall changes in relative humidity at the extreme dryend of the humidity range.

The consequences of drought in southeast Aus-tralia can be devastating, as was demonstrated bythe events of 2002. Within this context, the resultthat the LAPS model has a systematic tendency todrift away from the observed climate during ex-treme dry events is unsettling, and experiments areunder way to identify methods to improve thisrepresentation.

References

Council of Australian Governments, 2004. Report of the Na-tional Inquiry on Bushfire Mitigation and Management, byEllis, S., P. Kanowski, and R. Whelan, Australian GovernmentPublishing Service, Canberra.

Kanamitsu, M. and S. Saha, 1996. Systematic Tendency Errorin Budget Calculations, Monthly Weather Review, 124, 1145–1160.

Roads, J. O., M. Kanamitsu, and R. Stewart, 2002. CSE Waterand Energy Budgets in the NCEP-DOE Reanalysis II, Journalof Hydrometeorology, 3, 227–248.

Watkins, A. B., 2002. 2002 Australian Climate Summary: Dryand Warm Conditions Dominate, Bulletin of the AustralianMeteorological & Oceanographic Society, 15, 109-114.

Monthly water balance terms (mm/day) averaged overthe Murray-Darling Basin for March 2002 to February2003. Precip = precipitation, evap = evaporation (posi-tive upwards), dwdt = change in atmospheric water vapourstorage, divQ = water vapor flux divergence, res =residual, calculated as prec – evap + dw/dt + divQ.

May 200510

EUROPEAN PRECIPITATION EXTREMES:CAN WE EXPECT MORE?

O. B. Christensen1, J. H. Christensen1,and C. Frei2

1Danish Meteorological Institute2Atmospheric and Climate Science,

ETH-Zürich, Switzerland

In recent years Europe has been hit by severalsevere floods, like the Odra flooding in 1997 andthe Elbe and Rhone floodings in 2002. In thesearch for plausible expectations of future prob-abilities of this kind of events, a key issue is theevolution of precipitation extremes. Most globalclimate simulations will predict strong reductions insummer precipitation over central and southernEurope. However, the warmer climate will enablethe air to contain more water and should, every-thing else being equal, result in more severe extremeevents. An evaluation of how these two effects willbalance, requires quantitative calculations.

Global climate models (GCM), which have griddistances of hundreds of kilometers, are not wellsuited to directly deliver useful information aboutextreme precipitation, particularly, in smaller re-gions with complex terrain. One method to extractthe essential information from these GCMs is dy-namical downscaling, where regional climate models(RCM) with grid distances of 50 km or less pro-vide better spatial and temporal resolution.

Even though this strategy is computationallyexpensive, a recently finished European Union (EU)funded project, PRUDENCE, has managed to col-lect a large database of results from nine differentRCM models applied to a set of 30-year longexperiments covering past (1961–1990) and future(2071–2100) conditions. The canonical boundaryconditions that were used by all participating RCMsuse sea surface temperatures and sea ice extent forthe current conditions, and add a climate signal inthese quantities as calculated by the Hadley CentreGCM, HadCM3, for future conditions correspond-ing to the SRES A2 scenario. Global atmosphericinformation for these 30-year long time slices iscalculated with the help of the Hadley Centre high-resolution global atmospheric model HadAM3H.

The influence of the choice of a regional modelon the resulting climate is very dependent on the

season in question in Europe. Weather systemsduring winter are mainly determined by the large-scale circulation and hence predominantly determinedby the GCM delivering boundaries. In contrastthere is a large degree of variation between RCMsduring the summer, when local feedbacks are veryimportant.

Results from one participating model about Eu-ropean summer precipitation was reported inChristensen and Christensen (2003 and 2004), themain result being that the strong projected decreasein average precipitation did not prevent extremeprecipitation values like the 99th percentile of dailyprecipitation from increasing over most of Europe.In the top figure on page 20, which originates fromChristensen and Christensen (2004), results fromthe canonical PRUDENCE downscaling with theHIRHAM model in 50 km resolution is shown.Average precipitation and the exceedence of the99th percentile (roughly equivalent to the 1-yearreturn value) of daily precipitation are shown. Eventhough there are large reductions in average pre-cipitation, the extreme precipitation is generallyincreasing over most of Europe. The reduction ofaverage precipitation tracks a corresponding reduc-tion in the probability of precipitation; averageintensity changes only slightly in this simulation.

A recent analysis of several of the PRUDENCEexperiments yields some interesting results (Frei etal., 2005). All simulations yield reductions in aver-age precipitation of varying magnitude during summer.Extreme values, however, show positive changesfor some models but negative changes for others.In order to further describe the simulated changesin precipitation intensity probability density func-tion (PDF), Frei et al. (2005) fit a General ExtremeValue distribution to describe the extreme tail ofdaily precipitation during summer (JJA). Theycompare the changes of extremes simulated by theRCMs with a hypothetical change that would resultfrom a uniform scaling of the control PDF by thesimulated change in precipitation probability (fre-quency of wet days) and average precipitationintensity. The figure on page 11 shows results forCentral Europe and two different return values(5 and 50 years) in the form of a scatter plot ofsimulated change against a hypothetical (i.e., scaled)change. For all seven different RCMs the simulatedextremes increase more or decrease less than wouldbe expected from a uniform scaling of the controlPDF. There seems to be two conceptual compo-nents of the simulated change: One component,which is very coherent among the models, tends to

11May 2005

increase the occurrence of extremes. The secondcomponent, which expresses the effect of changesin precipitation probability and intensity partly com-pensates for the increase, but its magnitude variesconsiderably among the models.

Though this analysis still lacks a thorough un-derstanding of why the simulations behave thisway, there seems to be an overall agreement thatthere is an effect that enhances the PDF tail in thesimulations representing future conditions. There isstill a lot of work to do in order to determine moreexactly the relative changes of the two competingeffects. There is also an agreement that the climatechange signal will be more likely to be positive forlonger return periods.

Obviously, these results do not tell the wholestory about the changes in flood frequency thatcan be expected in the future. First of all, nosystematic validation of RCM performance withrespect to simulation of present-day extremes hasbeen performed on a continental scale. European

validation is currently limited to smaller areas likethe Alps (Frei et al., 2003, 2005) Denmark(Christensen et al., 2002) and the German state ofBaden-Württemberg (Semmler and Jacob, 2004).

Heavy precipitation is hard to measure accu-rately, and it would be beneficial to compare resultsof coupled-model simulations including both re-gional climate models and hydrological models toobserved river runoff. This is an interdisciplinaryfield of research where little has happened yet;results from the PRUDENCE project will be pub-lished as Graham et al. (2005). The futureimprovement of the present kind of climate projec-tions will require interdisciplinary work where bothhydrological and meteorological models are included.

References

Christensen, O. B., J. H. Christensen, and M. Botzet, 2002.Heavy precipitation occurrence in Scandinavia investigated witha Regional Climate Model, in Climatic Change: Implicationsfor the hydrological cycle and for water management, M. Beniston(ed.), Kluwer Academic Publishers.

Christensen, J. H. and O. B. Christensen, 2003. Severe sum-mertime flooding in Europe. Nature, 421, 805–806.

Christensen, O. B. and J. H. Christensen, 2004. Intensificationof extreme European summer precipitation in a warmer climate.Glob. Plan. Change, 44, 107–117.

Frei, C., R. Schöll, S. Fukutome, J. Schmidli, and P. L. Vidale,2005. Future Change of Precipitation Extremes in Europe: AnIntercomparison of Scenarios from Regional Climate Models.Submitted to J. Geoph. Res.

Frei, C., J. H. Christensen, M. Déqué, D. Jacob, R. G. Jones,and P. L. Vidale, 2003. Daily precipitation statistics in regionalclimate models: Evaluation and intercomparison for the Euro-pean Alps. J. Geophys. Res., 108 (D3), 4124, doi:10.1029/2002JD002287.

Graham, L. P., S. Hagemann, S. Jaun, and M. Beniston, 2005.On interpreting hydrological change from regional climate mod-els. Submitted to Climatic Change.

Semmler, T. and D. Jacob, 2004. Modeling extreme precipita-tion events – a climate change simulation for Europe. Glob.Plan. Change, 44, 119–127.

Simulated change in return levels (ratio scenario/con-trol) against the expected change from simple scalingof the control PDF by the changes in wet-day fre-quency and intensity for summer (JJA). Black symbolsare 5-year return levels, grey symbols are 50-yearreturn levels. Different symbols are differentdownscalings of PRUDENCE experiments. Both axesare logarithmic. From Frei et al. (2005).

INTERESTED IN SUBMITTING ANARTICLE TO BE PUBLISHED

IN GEWEX NEWS?

See the guidelines for submitting articles at:http://www.gewex.org/newsguidelines.htm

May 200512

U.S. HISTORIC DROUGHTCHARACTERISTICS ESTIMATED

USING N-LDASKonstantinos M. Andreadis1,

Dennis P. Lettenmaier1, Justin Sheffield2,and Eric F. Wood2

1University of Washington, Seattle, Washington,USA, 2Princeton University, New Jersey, USA

Drought is a pervasive natural hazard. Accord-ing to a recent report of the Western GovernorsAssociation (2003), an “extreme” or “severe” droughthas been experienced in some part of the UnitedStates in every year since 1895. Among U.S.natural hazards, drought is the most costly. A1995 report by the Federal Emergency ManagementAgency (FEMA, 1995) estimated that the annualcost of U.S. droughts was in the range $6-8B.The National Oceanic and AtmosphericAdministration's National Climatic Data Center(NCDC, 2003) estimates that since 1980, therehave been 10 droughts among the 57 natural disas-ters costing over $1B, and that the average annualcost to the U.S. of these $1B or larger droughtshas been almost $6B. The 1988 drought alone, in2002 dollars, cost almost $62B. This compareswith losses of about $27B for the 1993 MississippiRiver floods and about $36B for Hurricane An-drew. Among disasters with costs exceeding $1B,drought losses are about equal to the sum oflosses associated with floods and hurricanes, anddroughts are responsible for almost 40% of thetotal losses due to natural hazards (NCDC, 2003).

Methods for characterizing past droughts, andin turn the basis for evaluating ongoing droughts ina historic context, are not as well developed asthey are for other natural hazards, such as floods(e.g., estimation of flood return periods). A keyproblem is the spatial dimension of droughts. Whiledrought intensity and duration can be measured ata point, droughts are regional phenomena whosespatial characteristics vary over time. Nonetheless,the spatial extent of droughts is of equal, or greater,importance than drought intensity and duration.Many studies of drought have made use of thePalmer Drought Severity Index (PDSI) as a pointdrought indicator. However, inherent shortcomingsof PDSI (including not only its failure to charac-terize drought extent, but also other specifics suchas lack of cold land processes representation) havelimited its applicability. Other drought indicators,such as soil moisture in the case of agriculturaldrought or runoff as a measure of hydrologicdrought, provide a better physical basis for inter-

preting drought, however, they are constrained bydata availability. In particular, soil moisture multi-decadal data are virtually non-existent, whilestreamflow data are integrated over drainage areasthat are specific to particular gauges, and makeresolving the spatial variability of drought difficult.

An approach that overcomes these limitationsis the use of gridded land surface model simula-tions that are forced with carefully quality controlledclimatological data. These methods have beendeveloped in the context of the North AmericanLand Data Assimilation project (N-LDAS; Mitchellet al., 2004). Maurer et al. (2002) describe long-term retrospective land surface data sets developedas part of N-LDAS that include the main landsurface forcings (e.g., precipitation, temperature,and downward solar radiation), state variables (soilmoisture, snow water equivalent) and fluxes (evapo-transpiration, runoff). These data have been thebasis for studies of land-atmosphere teleconnections,hydrologic forecast initialization, and coupled land-atmosphere model initialization. The context issomewhat similar to land surface reanalysis prod-ucts (e.g., most recently from the European Centrefor Medium-Range Weather Forecasts (ECMWF)ERA-40 global reanalysis, and the NCEP NorthAmerican Regional Reanalysis (NARR), but a keydifference is that the surface forcings are based onobservations, and the model-predicted runoff (atleast at specified calibration points) is constrainedto match observations, at least approximately. Fur-thermore, by construct the surface energy and waterbalances are closed, hence at least for long-termaverages, evapotranspiration, and latent and sen-sible heat, should be correct. As for reanalysisproducts, the simulated land surface variables arespatially and temporally continuous. If the forcingdata sets are carefully quality controlled, the model-predicted variables (e.g., soil moisture and runoff)can be used to reconstruct drought histories. Thisspatial continuity allows for a more consistent wayof estimating the spatial extent of drought, whilethe temporal continuity and the resulting temporalresolution facilitates the capturing of short-termchanges of drought conditions, as well as longer-term variations.

The 50-year Maurer et al. (2002) data set wasused by Sheffield et al. (2004) to evaluate droughtevolution over the continental U.S. Seeking amethod that would allow direct comparison ofdrought severity over the conterminous U.S., theyestimated monthly soil moisture quantiles by fittingbeta distributions to the simulated soil moisturedata. The soil moisture quantiles were used as ameasure of drought occurrence and intensity [Vari-able Infilitration Capacity (VIC) drought index].

13May 2005

The use of quantiles is more appropriate thanabsolute soil moisture values, because anomalies ofabsolute magnitude reflect different drought severi-ties over different regions and cannot be used fordirect comparison of drought across the studydomain. Sheffield et al. (2004) identified a numberof dry events over the continental U.S. in the past50 years. The most notable included a multi-yeardrought in the southwest during the mid-1950s,two less severe events during the 1960s, a moder-ate drought over the western U.S. in 1977, andfinally a short-term drought in 1987–88 that cov-ered a large area of the study domain. On average,30% of the U.S. experienced severe droughts withthe peak occurring during the 1950s drought withover 70% of the domain covered by drought.

The regional characteristics of drought werealso examined, and it was found that the westernU.S. exhibited less variability and much higherpersistence than the eastern U.S. Comparison ofthe model-derived drought index with three differ-ent PDSI data sets for the same time period showedgood correspondence between the VIC and PDSI-based drought indices in terms of their year-to-yearvariability and trends. This can be seen in thefigure above which shows annual (top panel) andmonthly (lower panel) time series of the VIC droughtindex and PDSI averaged over the U.S. However,the limitations of the PDSI over mountainous re-gions were reflected in low correlations with theVIC drought index over the western U.S. (arguablythe VIC index is more realistic, given the model’srepresentation of snow-related process which domi-nate the hydrology of the region), while in the

eastern U.S. the correlation is much higher. Ingeneral, the simulated drought index was able tocapture most of the variability of soil moisture inthe context of drought identification over the 1950–1999 period.

Recently, NCDC has made station archivesavailable in electronic form from the beginning ofthe recorded record (in most cases, around 1900or later). Methods have been developed (see Hamletet al., 2005) for eliminating temporal heterogene-ities in the data sets by tying trends in the griddeddata to those in the smaller long-term station data(about 1200 stations) in the Historical ClimatologyNetwork (HCN). Essentially, the method adjuststhe gridded (daily precipitation and temperature)forcing data to have decadal-scale temporal vari-ability comparable to that of the HCN stations,while retaining the spatial variability of the NCDCstation data. As in the Maurer et al., 2002 dataset, other model forcings (downward solar andlongwave radiation, surface humidity) are estimatedusing algorithms that relate them to the daily tem-perature and/or temperature range. The NCEP-NCARreanalysis data, which Maurer et al., 2002 used forthe wind forcing, are absent prior to 1949, and sowe used average wind (varying by month) for theentire 1915–2003 period.

Andreadis et al. (2005) used the 1915–2003derived data (at one-half degree spatial resolution),following an approach similar to that used bySheffield et al., 2004 for drought identification—specifically, soil moisture and runoff were expressedas percentiles relative to the VIC climatology. Toinclude the spatial dimension of droughts, and toallow direct comparison of different droughts inthe long-term record, a method termed SeverityArea Duration (SAD) analysis was developed. SADsimultaneously captures the fundamental character-istics of drought (intensity, duration, and spatialextent). It is similar to Depth Area Duration analy-sis widely used for the estimation of precipitationfor design purposes, where precipitation depth isreplaced with an appropriate measure of droughtseverity.

Both agricultural and hydrological drought canbe examined using reconstructed long-term hydro-logic characteristics, in conjunction with the SADmethod. SAD facilitates direct comparison ofdroughts in the historic record that have varyingduration and spatial extent. The envelope of theSAD curves from all events allows examination ofthe most severe droughts for different durationsand areas. For instance, application of the SADmethod to the 1915–2003 reconstructed hydrologyof the continental U.S. shows that the droughts of

Comparison of annual (upper panel) and monthly (lowerpanel) time series of VIC drought index and three differ-ent PDSI data sets for 1950–99 averaged over theU.S. Figure from Sheffield et al. (2004).

May 200514

the 1930s and 1950s were the most severe for largeareas (interestingly, the “Dust Bowl” years of the1930s lie on the envelope curve for more intense,high severity events, but for shorter durations thanthe parts of the curves occupied by the 1950sdrought). Other notable events included the 1988Midwest drought, the 1977 drought over the west-ern U.S., and the mid 1960s drought over theeastern U.S. The 2000 western U.S. drought wasamong the most severe in the period, especially forshort durations and (relatively) small areas.

An interesting feature of the temporal evolutionof areas experiencing drought and drought severityis that the eastern U.S. generally exhibits muchlarger month-to-month variability than the centraland western U.S. which indicates the lower persis-tence of soil moisture. The bottom figure on page20 shows the spatial patterns of three major agri-cultural drought events of the 20th century alongwith the associated SAD curves for each drought(1930s, 1950s, and 2000s). Results for agriculturaland hydrological droughts were similar, althoughthe 2000 drought occupies a larger portion of thehydrological envelope SAD curve. This apparentlyis indicative of the faster response of runoff toprecipitation and the presence of short wet spellsduring the 1930s and 1950s but not during the2000 drought when precipitation anomalies exhib-ited longer persistence over the western U.S.

References

Andreadis, K. M., E. A. Clark, A. W. Wood, A. F. Hamlet,and D. P. Lettenmaier, 2005. 20th Century Drought in theConterminous United States, J. Hydrometeorology (accepted).

Federal Emergency Management Agency, 1995. National Miti-gation Strategy: Partnerships for Building Safer Communities,Federal Emergency Management Agency, Washington, D.C.,40 pp.

Hamlet A. F. and D. P. Lettenmaier, 2005. Production oftemporally consistent gridded precipitation and temperature fieldsfor the continental U.S., J. of Hydrometeorology (in press).

Maurer, E. P., A. W. Wood, J. C. Adam, D. P. Lettenmaier,and B. Nijssen, 2002. A long-term hydrologically-based data setof land surface fluxes and states for the conterminous UnitedStates, J. Climate, 15, 3237-3251.

Mitchell, K. E., D. Lohmann, P. R. Houser, and E. F. Wood,2004. The Multi-institution North American Land Data Assimi-lation System (NLDAS): Utilizing multiple GCIP products andpartners in a continental distributed hydrological modeling sys-tem. J. Geophys. Res. , 109, Art. No. D07S9010.1029/2003JD003823.

National Climatic Data Center, 2003. A Climatology of 1980-2003 Extreme Weather and Climate Events, Technical Report2003-1.

Sheffield J., G. Goteti, F. H. Wen, and E. F. Wood, 2004. Asimulated soil moisture based drought analysis for the UnitedStates. J. Geophysical Research, 109, Art. No. D24108.

GEWEX RELEVANT PUBLICATIONSOF INTEREST

Detectability of Anthropogenic Changes inAnnual Temperature and Precipitation ExtremesReference: Gabriele C. Hegerl, Francis W. Zwiers,Peter A. Stott, and Viatcheslav V. Kharin, 2004.Detectability of Anthropogenic Changes in AnnualTemperature and Precipitation Extremes. J. Climate,Vol. 17, No. 19, 3683–3700.Summary/Abstract: This paper discusses a study oftemperature and precipitation indices that may besuitable for the early detection of anthropogenic changein climatic extremes. Anthropogenic changes in dailyminimum and maximum temperatures and precipita-tion over land simulated with two differentatmosphere–ocean general circulation models are ana-lyzed. The use of data from two models helps toassess which changes might be robust between mod-els. Indices are calculated that scan the transitionfrom mean to extreme climate events within a year.Projected changes in temperature extremes are signifi-cantly different from changes in seasonal means overa large fraction (39–66%) of model grid points.Therefore, the detection of changes in seasonal meantemperature cannot be substituted for the detection ofchanges in extremes. Both models simulate extremeprecipitation changes that are stronger than the cor-responding changes in mean precipitation.

Terrain and Multiple-Scale Interactions as Factorsin Generating Extreme Precipitation Events

Reference: Roberto Rudari, Dara Entekhabi and GiorgioRoth, 2004. Terrain and Multiple-Scale Interactionsas Factors in Generating Extreme Precipitation Events.J. Hydrometeorology, Vol. 5, No. 3, 390–404.The Mediterranean region is often affected by floodingand landslides due to heavy precipitation events.The authors combine data from a dense network ofsurface precipitation gauges over northern Italy anda global atmospheric analysis at a coarser scale todevelop a multiscale diagnostic model of the phe-nomenon. Composite maps are formed based ondepartures from climatology and standard deviationof sea level pressure, 500-hPa geopotential, wind,and water vapor flux. A diagnostic model is builtbased on the evidence that shows the spawning ofsecondary mesoscale features in the steering syn-optic flow. The mesoscale features draw moistureand energy from local sources and cause extremeprecipitation events over adjoining areas. It isshown that longitudinal blocking frequency over alarger region strongly differs from climatology. Alow pressure center tracking algorithm is used tofollow the evolution of some events.

15May 2005

WORKSHOP/MEETING SUMMARIES

IGWCO/GEWEX/UNESCO WORKSHOPON TRENDS IN THE

GLOBAL WATER CYCLE

3–5 November 2004UNESCO, Paris, France

Rick Lawford1 and William Rossow2

1International GEWEX Project Office2NASA Goddard Institute for Space Science

More than 50 experts from the water cycle com-munity attended the Workshop, which was held toidentify significant advances in understanding trendsand to make recommendations that could be takenunder consideration by GEWEX, the Integrated Glo-bal Observing Strategy–Partners (IGOS-P) IntegratedGlobal Water Cycle Observations (IGWCO) Theme,and the Intergovernmental Panel on Climate Change.Presentations and discussions indicated that the as-sessment of climate trends from data is much morecomplex than it would first appear. For example,detection of climate change on a global scale requiresthat rigorous data standards be met. In particular,data sets must be global (to preclude misinterpreta-tion of regional changes as global changes), long-term(data records that are too short cannot distinguishbetween longer term variability and a true trend), andof sufficient accuracy (to distinguish the signal in thelarge volume of noise). It is also important to knowthe type of change for which one is looking.

Based on model projections associated with adoubling or a quadrupling of carbon dioxide, expertsexpect to see increases in surface temperatures, par-ticularly at higher latitudes over land areas. They alsoexpect to see an acceleration of water cycle pro-cesses on the order of 1 to 3% in evaporation andprecipitation over the last 50 years. Models indicatethat the relative humidity remains roughly constantproducing an increase of precipitable water at about7% per degree Celsius of warming.

In most parts of the world, the predicted trendsare relatively small in comparison to signals such asthe annual cycle. Given that the available data sys-tems were not designed to address long-term trenddetection and rarely are accompanied by publisheduncertainty estimates, few significant trends are indi-cated, and these are mainly confined to regionalrather than global manifestations. Workshop partici-pants unanimously accepted that there is evidencethat surface temperatures are increasing on a global

basis, but there were differences of opinion on someof the other global trends in shorter-term water cyclevariable records.

Regional trends are more frequently observed.For example, seasonal shifts in runoff timing (but notrunoff amounts) have been documented in midlatitudeareas especially where spring streamflows are depen-dent on snow melt. Only two to three decade timeperiod regional trends exist in precipitation patterns,although most of these tendencies currently are withinthe bounds of observational uncertainty. Naturalvariability is characterized by interannual variationsand longer-term (decadal to centennial) variations thatmust be separated from the effects of anthropogenicforcing. Several presentations demonstrated that datasets based on the current suite of satellites can beused to document and diagnose the causes of thisshorter-term variability and to better understand keyprocesses influencing the water cycle.

Another approach used in the past few years toidentify trends involves the use of model simulationsand reanalyses as tools to obtain trend estimates.Examples given at the Workshop included a studyshowing how an increase in lower stratospheric watervapor simulated with models is the same order as insitu balloon observations at Boulder, Colorado since1980. Other studies showed that there had been anincrease in specific humidity and changes in extremeprecipitation events. Although these techniques showtrends, they need to be confirmed with the analysisof data because the results from reanalysis systemsare affected by the quality of the input data sourcesand by the model physics.

Ongoing cross-comparisons of available data prod-ucts that are globally extensive and have recordsexceeding 1 to 2 decades suggest that the uncertain-ties in these data products can be significantly reduced.A comprehensive reanalysis of a combination of sat-ellite and in situ measurements should be undertakento obtain more homogeneous water cycle data prod-ucts for trend and variability analysis.

Preliminary conclusions from the workshop follow.

– While there is considerable confidence in thetrends derived from in situ data records fortemperature, the same confidence does notexist for water cycle variables (i.e, precipi-tation or clouds).

– Some of the strongest trends seen in hydro-logic variabilities (glacier retreat, changes inthe seasonality of runoff) are to a largeextent, responses to global warming ratherthan responses to changes in the global watercycle variables (i.e, precipitation).

May 200516

The Fourth International Implementation/SciencePlanning Meeting for the Coordinated EnhancedObserving Period (CEOP) and the First IntegratedGlobal Observing Strategy Partners (IGOS-P) Inte-grated Global Water Cycle Observations Theme(IGWCO) Implementation Workshop, was held atthe Sanjo Kaikan facility on the Hongo Campus, ofthe University of Tokyo. The meeting was heldjointly in recognition of the fact that CEOP hadearlier been endorsed as the first element of theIGWCO theme within the framework of IGOS-P.The CEOP Lead Scientist, Toshio Koike, reportedon actions taken in 2004 to strengthen the connec-tions between the two communities includingdeveloping a plan for the second phase of CEOPduring this joint planning event.

The participants addressed several importantissues including: (a) a concept for finalizing theCEOP Phase 2 Implementation Plan; (b) maximiz-ing the science and technology benefits from bothCEOP and IGWCO; and (c) a framework foroversight of the science, implementation plans andresults during the initial phase of IGWCO andCEOP Phase 2. The joint meeting also resulted ina number of specific recommendations for efficientorganization and management of both activities toensure they would achieve their main objectives.The scope of the initial implementation of IGWCOand the second phase of CEOP will be coordi-nated with the needs of the World Climate ResearchProgramme Coordinated Observation and Predic-tion of the Earth System (COPES) framing strategyand connections to the 10-year implementation planfor a comprehensive, coordinated, and sustainedGlobal Earth Observation System of Systems(GEOSS).

This meeting included a CEOP Phase I Scienceand Data Results Workshop with 42 technical pa-pers and a similar number of posters. Participantscovered issues associated with model and satelliteinstrument algorithm validation and intercomparisonwork; water and energy budget variations and theirrole in climate; monsoon characteristics studies,including diurnal variations, transferability of model

FOURTH CEOP PLANNING MEETINGAND FIRST IGWCO WORKSHOP

28 February – 4 March 2005Tokyo, Japan

Sam BenedictInternational CEOP Coordinator

– Current reanalysis products from models donot provide an adequate basis for identifyingtrends in climate variables due to uncertaintiesin data, including inputs, model physics, andresolution.

– While global maps derived from satellite dataare reaching the point where they have longenough time series for confirming trends, manyof these products have not been optimized forhomogeneity of record and for climate analy-sis. These satellite products need to bereprocessed on an urgent basis to provide areliable basis for future assessment of recentclimate trends.

Specific recommendations (partial list) include:

1. An assessment of the utility of data sets forclimate change detection should be under-taken by identifying limitations and artifactsof the analysis procedures.

2. The global climate modeling community shouldprovide “fingerprints” of the changes in watercycle variables expected as the climate changes.

3. A better process for bringing the observa-tional community together with the climatediagnostics community is needed to improvedata sets.

4. Data providers should have a more consis-tent approach to the use of radiance data.

5. A strategy is needed for dealing with theheterogeneity of in situ and satellite data andfor merging these data to generate long-termdata sets.

6. An analysis should be carried out to assessthe consistency of trends in the full range ofwater cycle variables.

7. The Earth Observation Summit ad-hoc Groupon Earth Observations (GEO), IGOS-P, andWMO need to encourage the developmentand acceptance by nations of common datastandards and data sharing practices. Thesestandards also should address regional radarprecipitation data sets.

8. The global, long-term data sets, including olderversions of these data (by data rescue) shouldbe evaluated (by cross-comparison and inves-tigation of differences) and reprocessed.

9. Satellite agencies should make long-term com-mitments to producing principal climate dataproducts.

17May 2005

results and downscaling. Extended abstracts ofthese presentations have been published as a sepa-rate meeting document and can be viewed atwww.ceop.net. Many of the studies used theCEOP data sets, while others were examples of thetype of research and data handling that need CEOP’sunique capabilities to succeed. Highlights of thesecapabilities include:

– A prototype of the global water cycle obser-vation system of systems based on thereference site network, the experimental andoperational satellite systems, and the numeri-cal weather prediction model outputs.

– A well organized data archive system that wasestablished in cooperation with the World DataCenter for Climate at Max Planck Institute forMeteorology in Hamburg, Germany; the Uni-versity Cooperation for Atmospheric Researchin Colorado, U.S.A.; the University of Tokyoin Japan; and the Japan Aerospace ExplorationAgency in Tokyo, Japan.

– A cooperative framework for providing dis-tributed-and centralized-data integrationfunctions for use by the scientific and re-mote sensing communities. CEOP internationalinformation sharing and dissemination stan-dards that are compliant with existing standardsand capabilities, such as the InternationalOrganization for Standardization (ISO), andother appropriate technologies.

The discussions also focused on the links be-tween IGWCO and other IGOS themes with theGlobal Water System Project (GWSP). In particu-lar the data needs of GWSP were examined alongwith a possible prototype application of the ICSUpaper on socio-economic data. The status ofplans for an IGWCO related indicator workshop,findings and the status of follow-up to earlierworkshops on indicators, remote sensing and waterquality, and specific plans within IGWCO for workin this area were reviewed. Several cross-cuttingtopics related to improved integration of IGOS-Pthemes were discussed including disasters (e.g.,tsunamis) and climate. Opportunities for capacitybuilding activities were presented and reviewed byrepresentatives from various countries.

The actions and recommendations related tothese discussions are being drafted into a report ofthe joint meeting. The next CEOP meeting will becoordinated with IGWCO and will also includerepresentatives of COPES. It is scheduled for27 February to 3 March 2006 in Paris, France.

TWENTY-SIXTH JSC MEETING

14–18 March 2004Guayaquil, Ecuador

Rick LawfordInternational GEWEX Project Office

The meeting of the Joint Scientific Committee(JSC) for the World Climate Research Programme(WCRP) was held at the Escuela Superior Politechnicadel Litoral Center and was hosted by Dr. PilarCornejo and her colleagues. A primary issue fordiscussion at the meeting was the CoordinatedObservation and Prediction of the Earth System(COPES) strategy and its implementation. The reportof the framing task force was reviewed and otherreports were given by the COPES Task Force andPanels. The Task Force on Seasonal Predictionoutlined its plans for an interactive modeling systemfor prediction studies. The role of COPES as anintegrating framework for WCRP was adopted andits relationship to existing projects is being clarified.

Three priority issues were discussed as an earlyimplementation of COPES: sea level rise, anthropo-genic climate change (ACC), and monsoons. Thesea level rise discussion focused on plans for aworkshop dealing with the factors contributing tosea level rise. Issues such as the development ofan ability to remotely measure changes in the quan-tity of water stored on land are important for thesediscussions. Progress in understanding ACC iscritical to achieving WCRP’s objectives and aneffective initiative is needed to address this issue.After presentations on monsoon activities in theClimate Variability (CLIVAR) Project and GEWEX,the JSC encouraged closer collaboration in themonsoon work research underway in the variousparts of WCRP. The Pan-WCRP Monsoon Work-shop planned for 15–17 June 2005 should providea basis for this collaboration.

Based on a self-assessment of its activities,CLIVAR is reorganizing into four major themes: ElNiño/Southern Oscillation, ACC, monsoons, andthermohaline circulation/decadal. The Cryosphereand Climate (CliC) Project is broadening its link-ages with programs outside WCRP and uncoveringsome new science issues such as uncertainties inthe lags between the climate forcing of thecryospheric system and the system’s response.The Stratospheric Processes and their Role in Cli-mate (SPARC) Project reviewed their progress overthe past year and provided rationale for collabora-

May 200518

tive studies with other WCRP projects, includingwork with GEWEX in using cloud resolving mod-els to address questions related to the dehydrationof the upper troposphere in areas around clouds,and in assessing aerosol distributions and the abil-ity of the radiation codes in climate models andsatellite algorithms to account for the wide varietyof aerosol types and measurements.

GEWEX highlighted its progress in incorporat-ing land characteristics into seasonal predictionmodels; new initiatives related to cloud processstudies in support of climate model development;new data products including the International Sat-ellite Land Surface Climatology Project (ISLSCP)Initiative II data sets, and the adoption of theBaseline Surface Radiation Network (BSRN) as abaseline network for the Global Climate ObservingSystem (GCOS). Modifications to the GEWEXPhase II objectives were presented and approved.The JSC will review the Coordinated EnhancedObserving Period (CEOP) Phase II Science andImplementation Plan and has recommended thatCEOP report on observational issues to the WCRPObservations and Assimilation Panel.

A number of projects associated with WCRPreported on their progress. GCOS outlined its re-search needs for support to observing programsbeing enhanced in its implementation plan. Morethan 350 preproposals have been received for theInternational Polar Year covering topics related toArctic change, global linkages, human dimensions,and data sets. An overview of The ObservingSystem Research and Predictability Experiment(THORPEX) emphasized the longer term weatherprediction research that has interest for WCRP.Planning is underway to increase collaboration be-tween WCRP and the International Geosphere-Biosphere Project (IGBP). GEWEX has beenpaired with the IGBP Integrated Land Ecosystem–Atmosphere Processes Study as part of this review.The Global Water System Project has adopted anumber of fast track projects including the prepa-ration of an electronic “water atlas” and participationin the Northern Eurasia Earth System ProcessesInitiative. WCRP efforts that are contributing ex-tensively to the Earth Observation Summit ad-hocGroup on Earth Observations, the preparation ofan Integrated Global Observing Strategy CryosphericTheme and the implementation of the IntegratedGlobal Water Cycle Observations Theme.

The JSC meeting marked a major step in thedevelopment of COPES and increased collabora-tion between WCRP projects.

WORKSHOP ON THE APPLICABILITYOF CLIMATE RESEARCH FOR WATER

RESOURCE MANAGEMENT INSEMI-ARID AND ARID REGIONS

18–20 April 2005Cairo, Egypt

Lawrence MartzUniversity of Saskatchewan

Saskatoon, Canada

In response to a request from the Arab statesto the World Meteorological Organization, a spe-cial Workshop was held to examine the applicabilityof climate research and information for water re-source management in semi-arid and arid regions.The workshop was hosted by the World ClimateResearch Programme (WCRP) and the United Na-tions Educational Scientific and Cultural Organization(UNESCO), and organized by the GEWEX WaterResources Applications Project (WRAP), and theUNESCO-International Hydrological Programme (IHP)Cairo Office.

The program opened with an official welcomeon behalf of the Egyptian Ministry of Water Re-sources and Irrigation, an introduction from theco-chairs Radwan Al-Weshah (UNESCO) andLawrence Martz (GEWEX-WRAP), and a welcomefrom Gilles Sommeria on behalf of WCRP.Mohamed Abdulrazzak, Director of the IHP CairoOffice, and Abdin Salih, Director of the IHP TehranOffice actively participated in the workshop.

The relationship between the Earth climate sys-tem and the hydrology and water resources of aridand semi-arid regions of the world is of particularinterest to the decision makers and stakeholders inthese regions. With the projected population growthin the semi-arid regions and the increasing demandfor adequate water resources, these stakeholderswish to know information that is useful for waterresources management.

The workshop addressed this overarching issueby bringing together operational hydrology and wa-ter management stakeholders from the arid/semi-aridregions of North Africa and the Middle East withhydro climate scientists engaged in observation,modeling, and analysis. Presentations includeddescriptions of regional water issues and examinedthe current state of knowledge about the climatesystem relevant to water management in arid/semi-

19May 2005

arid regions. Specific regional presentations weremade on the major water issues in the region,including representatives from Lebanon, Jordan,Sudan, Syria, the United Arab Emirates, Egypt andIran.

Gaps in understanding, data availability andregional capacity were identified in reviews ofhydroclimatological data products and modellingachievements, including activities undertaken underthe Water and Development Information for AridLands-A Global Network (G-WADI) Initiative.

The principal actions identified in the workshopdiscussions include:

• Increase awareness of the significance ofclimate change and variability in water re-source planning and management across theregion;

• Improve sharing of hydroclimatological dataacross various national and intra-national ju-risdictions;

• Access and develop tools to bring globaland regional data into a form suitable tosupport decision-making;

• Apply climate forecasts and data productsto specific water management and planningissues; and

• Improve the interface between regional insti-tutions and the international community, inorder to enhance the regional capacity toaddress the above issues.

The Workshop participants recommended thaton a priority basis, a regional network of profes-sional and academic scientists be formed. Existinginstitutional capacity within the region should beused as much as possible to develop and supportsuch a network and UNESCO was identified ashaving a lead role in this regard. The initial tasksof the network would be to: (1) inventory expertisewithin the region; (2) inventory of hydroclimatologicalobservation sites across the region; (3) examine thefeasibility of developing a reference site under theCEOP program; and (4) develop a proposal for ademonstration project that applies GEWEX/CLIVARdata products or models to the solution of aspecific water management issue.

The workshop established a foundation for fu-ture enhanced collaboration and the application ofWCRP products to the solution of pressing watermanagement problems.

GEWEX/WCRP MEETINGS CALENDARFor a complete listing of meetings, see the

GEWEX web site (http://www.gewex.org)26 May 2005—IGOS PARTNERS MEETING, Geneva,Switzerland.1–3 June 2005—WCRP OBSERVATION AND ASSIMI-LATION PANEL, New York, USA.13–17 June 2005—AMS JOINT CONFERENCES ONATMOSPHERIC AND OCEAN FLUID DYNAMICS,MIDDLE ATMOSPHERES AND CLIMATE PREDICT-ABILITY, Cambridge, Massachusetts, USA.

15–17 June 2005—PAN-WCRP MONSOON WORK-SHOP, Irvine, California, USA.

20–24 June 2005—5TH INTERNATIONAL SCIENTIFICCONFERENCE ON THE GLOBAL ENERGY ANDWATER CYCLE, Costa Mesa, California, USA.

19–22 July 2005—2005 WATERSHED MANAGEMENTCONFERENCE, Williamsburg, Virginia, USA.

26–28 July 2005—1ST INTERNATIONAL SYMPOSIUMON TERRESTRIAL AND CLIMATE CHANGE INMONGOLIA, Ulaanbaatar, Mongolia.

2–11 August 2005—IAMAS SCIENTIFIC ASSEMBLY,Beijing, China.

24–26 August 2005—GEWEX EXECUTIVE MEETING,Maryland, USA.

29 August – 1 September 2005—IGU CONFERENCEON ENVIRONMENTAL CHANGE AND RATIONAL USEOF WATER, Buenos Aires, Argentina.

14–16 September 2005—3RD WCRP OFFICERS,CHAIRS AND DIRECTORS MEETING, Geneva,Switzerland.

19–21 September 2005—JOINT GABLS/GLASS WORK-SHOP ON LOCAL COUPLED LAND-ATMOSPHEREMODELLING, De Bilt, The Netherlands.

19–23 September 2005—LOCO/GABLS WORKSHOP,De Bilt, The Netherlands.

21–23 September 2005—6TH GLASS PANEL MEET-ING, De Bilt, The Netherlands.

26–30 September 2005—11TH MEETING OF THEGEWEX HYDROMETEOROLOGY PANEL, Melbourne,Australia.

3–6 October 2005—16TH SESSION OF THE GEWEXRADIATION PANEL, Paris, France.

3–5 October 2005—9TH SESSION OF WGCM ANDJOINT WGCM-WMP MEETING, Exeter, UK.

6–7 October 2005—1ST SESSION OF THE WCRPMODELLING PANEL, Exeter, UK.

May 200520

Relative precipitation change in percent between control and scenario for the HIRHAM downscaling of the HadAM3Hsimulation during late summer (JAS). Left panel: average precipitation. Right panel: exceedence of the 99th percentile.From Christensen and Christensen (2004). See article on page 10.

Maps of spatial patterns of 3 major droughts (1930s, 1950s, 2000s). Each map corresponds to the month of highestseverity and spatial extent (upper panel). The lower panel shows the Severity Area Duration (SAD) analysis curvesfor the respective droughts, for selected durations. See article on page 12.

GEWEX NEWSPublished by the International GEWEX Project Office (IGPO)

Richard G. Lawford, DirectorDawn P. Erlich, Editor

International GEWEX Project Office1010 Wayne Avenue, Suite 450Silver Spring, MD 20910, USA

Tel: (301) 565-8345Fax: (301) 565-8279E-mail: gewex@gewex.orgWeb Site: http://www.gewex.org

SPATIAL DISTRIBUTION OF THREE MAJOR US DROUGHTS

RESULTS FROM "PRUDENCE" EXPERIMENT SHOW EXTREMEPRECIPITATION INCREASING OVER EUROPE