SUMMARYSUMMARYSUMMARYSUMMARY REVIEWREVIEWREVIEWREVIEW OFOFOFOF THETHETHETHE SHANGHAISHANGHAISHANGHAISHANGHAI 2010201020102010 WORLDWORLDWORLDWORLD EXPOEXPOEXPOEXPO NOWCASTINGNOWCASTINGNOWCASTINGNOWCASTING SERVICESSERVICESSERVICESSERVICES(WENS)(WENS)(WENS)(WENS) DEMONSTRATIONDEMONSTRATIONDEMONSTRATIONDEMONSTRATION PROJECTPROJECTPROJECTPROJECT

Jianhua DAI and Lan TAOShanghai Meteorological Service of China Meteorological Administration, Shanghai, China 200030

1.1.1.1. INTRODUCTIONINTRODUCTIONINTRODUCTIONINTRODUCTION1.11.11.11.1WeatherWeatherWeatherWeather serviceserviceserviceservice requirementsrequirementsrequirementsrequirements ofofofof ExpoExpoExpoExpo

The Shanghai World Expo 2010 was heldfrom May to October in 2010 (184 days). Underthe theme of "Better City, Better Life", the Expo puton display the splendid achievements ofcontemporary civilizations, pooled human wisdomin exploring the path of urban development, setmany Expo-related records, and added a gloriouschapter to the Expo history. This successfulinternational event attracted 246 officialparticipants, including 190 countries and 56international organizations, and 73 million visitors.All these figures are new records in World Expohistory, the first of which was held in London in1851.

Fig1 Closing Ceremony of the Shanghai World Expo 2010 on

Oct. 31, 2010

With a pleasant northern subtropical maritimemonsoon climate, Shanghai enjoys four distinctseasons, with generous sunshine and abundantrainfall. The city is bordered by the Yangtze Riverto the north and Hangzhou Bay to the south.Shanghai receives an average annual rainfall of1,200 mm; nearly 60% of the precipitation occurs

from April to September. July and August areShanghai's hottest months with average highs of32.4°C, and daytime average of humidity is 75%.On average there were 15 rainy days, and 8thunderstorm days per month during the Expoperiod of operations. Abnormal rainfall events canoccur during this period, e.g. on the 25th of August,2008, the precipitation exceeded 100 mm in a onehour period.

Expo brought a lot of requirements ofmeteorological services. Planning an event duringthe warm season of a year in Shanghai, a time offrequent precipitation and thunderstorms, togetherwith some high impact weather (HIW), such asheavy rainfall, squalls, hail, lightning, and tropicalcyclones could bring weather-related disasters. Inaddition, Expo 2010 Shanghai was the first Expoto be held in the downtown area of a mega-city, theExpo site was located on the banks of theHuangpu River in the downtown area, covering anarea of about 5.3 km2, which is a relatively smallarea or a point in terms of operational weatherforecasting. The topography, surface features, andthe urban “heat island” effect could play importantroles in the initiation and evolution of weathersystems, which increases the difficulty offorecasting meso-scale weather.

Weather services were also required tosupport planning and coordination of EXPOoperations and activities. Important issuesconcerning the weather’s effect on individualexhibition pavilions and potential impactsassociated with the outdoor exposure of a largenumber of people, and a large extent of property.

There were about 250 pavilions in the Expo site,some important activities (including the OpeningCeremony, the Closing Ceremony, parades, etc.)and more than 20,000 smaller scale activities, over70 million visitors during the Expo period, whichdemanded detailed, targeted, and tailoredmeteorological services.

The Shanghai Meteorological Service (SMS)of China Meteorological Administration (CMA) wasresponsible for providing weather services for thewhole Expo activity. Challenges of themeteorological services included the difficulty offorecasting these frequent high impact weatherduring the Expo period.

Nowcasting is defined as Very Short-Range(0-6 hour) forecasts, and nowcasts of HIW areparticularly important components of the WorldEXPO 2010 weather service requirement. In caseof thunderstorms and other severe weatherforecast, severe weather outlooks (0-12 hour)need to be normally issued at 6 hour intervals forthe EXPO site, and special early warninginformation need to be issued through theDissemination Platform of the ShanghaiMulti-Hazards Early Warning System (MHEWS) ina timely manner.

In November 2008, the World ExpoNowcasting Services (WENS) project was officiallyestablished as one of the World MeteorologicalOrganization’s (WMO) demonstration projects.The WENS Implementation Plan was draftedthrough the joint efforts of the experts from WMO,China Meteorological Administration (CMA),Australia Bureau of Meteorology (BOM), HongKong Observatory (HKO), and CMA’s SMS andBeijing Meteorological Service (BMS).

At the recommendation of the WMOImplementation and Coordination Team on PublicWeather Services, WENS focused on the

"services" rather than the "systems", in particular,the translation of nowcasting system outputs totimely, relevant and user-friendly products fordecision makers and stakeholders. WENSproducts were made available to local weatherforecasters on duty. A post-project review wasconducted to assess the impact of WENS. Thiswas to be followed by the publication of guidelineson the provision of nowcasting services reflectingthe experience gained from WENS, and theconduct of capacity building workshops for WMOMembers. This project was conducted in theframework of the “Learning through Doing”concept as developed by PWS ProgrammeExperts.

The WENS has been conducted over fouryears (2008-2011) by CMA and internationalparticipating groups with active support from WMO.The formal WENS was conducted over the period1 March to 31 October 2010. This project wasalso in the framework of the “Learning throughDoing” concept as developed by PWS ProgrammeExperts.

1.21.21.21.2 GoalsGoalsGoalsGoals andandandand objectivesobjectivesobjectivesobjectives ofofofof WENSWENSWENSWENS

The World EXPO presents a unique andexciting opportunity to reach out to the generalpublic on a global level. The WENS aimed to be anexcellent example of how Meteorological andHydrological Services can participate ininternational events in more than just a servicecapacity. The improvements in DRR (Disaster RiskReduction) for natural disasters as well climatechange adaptation strategies are highlighted andpublic awareness of meteorological sciences andservices are gradually increasing.

Therefore, the goals of WENS are as follows:• In the context of multi-hazard early warningservices (MHEWS), to demonstrate howNowcasting applications can enhance short-range

forecasts of severe weather using the opportunityafforded by the Shanghai World EXPO 2010; and

• Promote the understanding and enhance thecapability, as appropriate, of WMO Members innowcasting services.

The objectives of WENS are as follows:• Provide advanced high impact weather andprecipitation nowcasting products and services inthe context of the World EXPO 2010;

• Enhance the capacity of the SMS in MHEWSto: Address the problem of urban inundation;Provide improved heavy precipitation warnings;Evaluate the contribution of QuantitativePrecipitation Estimates (QPE) and QuantitativePrecipitation Forecasts (QPF) to the overalleffectiveness in the risk assessment process;Effectively present the information to the decisionmakers and the public.

• Demonstrate the introduction, optimalimplementation, and training in use (technologytransfer) of advanced nowcasting systems inoperational forecasting and in the generation ofenhanced products and services;

• Evaluate the impact of the implementation ofoperationally focused nowcasting on the quality ofhigh impact weather and precipitation forecasts,on forecasters and on end-users of a localmeteorological service;

• Promote the implementation of nowcastingservices in the Shanghai region initially andultimately for the benefit of WMO Members,especially those in East Asia.

The nowcasting service product providers forWENS project are China MeteorologicalAdministration (CMA), Australian Bureau ofMeteorology (BOM), and Hong Kong Observatory(HKO). The participating nowcasting systems are:

The Terms of Reference for NSPPs are as

follows:

a) To provide nowcasting services products orsystems in support of the World EXPO 2010on a functional real-time basis;

b) To assist in the customisation, support,maintenance and interfacing of their systemsto operational data feeds and information /dissemination systems as agreed forimplementation in Shanghai;

c) Offer consultancy and advisory support andundertake agreed interface andcapacity-building activities.

1.31.31.31.3 OrganizationOrganizationOrganizationOrganization ofofofof WENSWENSWENSWENSWMO/WMO/WMO/WMO/

PWSPPWSPPWSPPWSP/WWRP/WWRP/WWRP/WWRPCMACMACMACMA

WENS Science Steering WENS Science Steering WENS Science Steering WENS Science Steering Group Group Group Group

WENS Working WENS Working WENS Working WENS Working Group Group Group Group

WENSWENSWENSWENSParticipantsParticipantsParticipantsParticipants

WENS WENS WENS WENS LogisticalLogisticalLogisticalLogistical

Support GroupSupport GroupSupport GroupSupport Group

WMO/WMO/WMO/WMO/PWSPPWSPPWSPPWSP/WWRP/WWRP/WWRP/WWRP

CMACMACMACMA

WENS Science Steering WENS Science Steering WENS Science Steering WENS Science Steering Group Group Group Group

WENS Working WENS Working WENS Working WENS Working Group Group Group Group

WENSWENSWENSWENSParticipantsParticipantsParticipantsParticipants

WENS WENS WENS WENS LogisticalLogisticalLogisticalLogistical

Support GroupSupport GroupSupport GroupSupport Group

Fig 2 Organization of WENS

A WENS Scientific Steering Group (SSG) wasestablished to plan and steer implementation ofthe project recognizing a requirement for a focuson service delivery aspects. The SSG is in chargeof developing an agreed business/project planbased on the MHEWS project plan. The WENSproject plan defines project participant roles andresponsibilities, agreed timetable and financialarrangements. A WENS Working Group(WENSWG) oversees the development of animplementation plan, its execution and reports tothe SSG.

The responsibilities of the WENS/SSG are: toformulate strategies of the WENS implementation(scientific and technical aspects), to monitor

progress of WENS implementation includingreview of the implementation plans of theparticipants, to oversee training programme, andprepare reports on scientific aspects of WENS.

The SSG comprises SMS/CMA, PWSSecretariat, ICT Chair, JONAS SC Co-chairs andPSOs, capacity building lead, a socio-economiclead with the WENSWG Chair serving as secretary.The WENSWG comprises of SMS/CMA experts,PSO representatives, PWS expert(s) and otherexperts deemed necessary serving as resourceagents.

A comprehensive user evaluation processwas also developed and applied to all WENSproducts and services by the WENS WorkingGroup.

1.41.41.41.4 MilestonesMilestonesMilestonesMilestones ofofofof WENSWENSWENSWENS

From the beginning of 2009, WENS relatedwork was pushed forward smoothly in accordancewith the Implementation Plan thanks to the greatefforts and contributions of the participant systems.There are milestones of the whole WENS project:

1) Sep. 2008: Establishment of WENS ScientificSteering Group (SSG) and WENSWG by interimSSG, draft TORs and draft business plan forWENS. (Interim SSG).

2) Nov. 2008: SSG Planning Meeting finalizedTORs and WENS Business Plan on project scope,duration, milestones and deliverables as well asfunding arrangements (SSG).

3) NEED Date: Pre-Training in Hongkong andAustralia: HKO (Linus Yeung and his colleagues)and BOM (Alan Seed and his colleagues) providedtrainings to the SMS staff on application of thenowcast products of SWIRLS and STEPS.

4) Jul.-Aug. 2009: First trial run (WENSWG): the

participant systems were installed and fine-tunedunder the guidance and support of the experts andoperational training courses for WENS forecasterswere taken.

5) Oct. 2009: Interim Review Meeting: Reviewobjectives, undertake validation and performanceassessment of systems taking part in trial run;engage key stakeholders/users to providefeedback on products.

6) Apr. 26 – Oct. 31, 2010: WENS in FullOperation.

7) Sep. 20-24, 2010: WENS Reviewpresentation, in the WMO Public Weather ServicesCore Implementation and Co-Ordination Teammeeting, Shanghai, China.

8) 8-10 Feb., 2011: WENS Review presentationon internet, in the WMO WWRP Working Group onNowcasting Research, Geneva.

9) Nov. 14-18, 2011: WENS Final Reviewmeeting of WENS was held in Shanghai, China.

2.2.2.2. DATADATADATADATA ANDANDANDAND WEBWEBWEBWEB SITESSITESSITESSITES2.12.12.12.1 DataDataDataData

There are 200 automated weather stations(AWSs) or rain gauges in the Shanghai area. Thespatial resolutions of the AWS network and therain gauge network are about 8 km and 5 km,respectively. The temporal resolution of thesetwo observation systems is 1 min. All the dataare communicated in real-time using theGPRS/CDMA.

There are two Doppler weather radars inShanghai, a WSR-88D Doppler weather radar inthe east coastal area of Shanghai, and a CINRADin the western area of Shanghai. They run avolume scan each 6 min and provide three typesof base data: reflectivity, radial velocity, and

spectrum width, and more than 70 derived radarproducts. With the CINRADs at Zhoushan, Ningbo,Changzhou, Nantong, and Hangzhou, a Dopplerweather radar network over the Yangtze RiverDelta provides weather information in real time.

A total lightning detection network detectstotal lightning that includes Intra-Cloud (IC)lightning and Cloud to Ground (CG) lightning. Thelightning data are transported with the CDMAtechnique to the central station in real time.

Doppler radar volume scan level-2 data,geostationary satellite data, conventional climatestation and radiosonde data, all real-time AWS andlightning location data in Shanghai, Jiangsu, andZhejiang.

2.2.2.2.2222 WebWebWebWeb sitesitesitesite

Fig 3 Professional WENS web site (left) in SMS and the public

WENS web page (right)

An official WENS web site was set up in theSMS’s intranet for professional users at Expo.These professionals included forecasters in theSMS weather office and in the Expo OperationCenter, and service officers and coordinators ofthe SMS. On the page, observations andnowcasting products such as QPE, QPF, SevereStorm Nowcasts, and a host of other participatingsystems could be displayed.

A temporary web page on the public web siteof the SMS displaying real time WENS productswas available for the public during the WENSperiod. Some key users and specialized users hadtheir own accounts to visit their own pages andproducts.

3. PARTICIPATING NOWCASTINGSYSTESMS

Four CMA nowcasting systems participatedthe WENS project: SWAN (Severe WeatherAutomatic Nowcast system), BJANC (BeijingAuto-Nowcastor), NoCAWS (NowCAsting andWarning System), and SMS-WARR (WRFADAS-3DVar Rapid Refresh system). TheAustralia BOM provided the STEPS (Short TermEnsemble Prediction System), and HKO SWIRLS(Short-range Warning of Intense Rainstorms inLocalized Systems).

3.13.13.13.1 BJANCBJANCBJANCBJANC (CMA-BMS)(CMA-BMS)(CMA-BMS)(CMA-BMS)

BJANC is a fruit of an internationalcollaborative project between China (BMS) andUSA (NCAR). BJANC was focused on severeconvective storm analysis and nowcast in supportof the 2008 Olympics, and convective weatherwarning in Beijing area. BJANC produces 0-1 hrforecasts of convective storm location andintensity with optional human enhancement toforecasts by entering boundary layer convergenceline or modifying focused forecast areas. Severestorm extrapolation from the system can beextended to 2 hrs. BJANC combinesmeteorological observations, high-resolutionlow-level thermodynamic structure analysis basedon a 4D-Var technique, rapid-updating Meso-NWPoutputs and several storm feature detection andtracking algorithms, as well as local data analysisalgorithms, to provide routine nowcasts of position,intensification and evolution of convective storm.The resultant nowcasts of storm initiation, growth,

and decay can be used as guidance by theforecasters or nowcasters.

3.23.23.23.2 NoCAWSNoCAWSNoCAWSNoCAWS (CMA-SMS)(CMA-SMS)(CMA-SMS)(CMA-SMS)

NoCAWS is an operational system for severeweather very-short range forecasting andnowcasting, with the following functions: dataprocess and management, GIS-based integrateddisplay, severe weather automatic alert,nowcasting products of precipitation and lightningstroke, analysis tools (sounding and 3-D lightning),and interactive nowcasting and warninggeneration. NoCAWS uses some relevanttechniques in the system, such as severe weatheroutlook, regional severe weather watch radar echoextrapolation, QPE, QPF, storm life-cycle analysis& estimate, and lightning threat prediction. UsingNoCAWS, the local forecasters can monitorweather, make mesoscale weather analyses,analyze storm environments, diagnose stormstructures and life cycles, and interactively makenowcasting and warning products. Some serviceproducts can be generated automatically orinteractively by the forecasters. The serviceproducts are the severe weather outlook (0-12h)for the Yangtze River Delta (YRD) region andShanghai city, severe weather watch (0-6h) for theYRD region, severe weather warning (0-2h) forShanghai city and end users.

3.33.33.33.3 STEPSSTEPSSTEPSSTEPS (BOM)(BOM)(BOM)(BOM)

STEPS generates ensembles of rainfallforecasts in the 0-6 hour lead time by blending theradar based extrapolation nowcast with theNumerical Weather Prediction (NWP) rainfallforecasts and then perturbing the blend with astatistical model of the forecast error. The life timeof a storm generally increases with the size of thestorm, therefore large rain areas in the radarimage can be predicted with greater accuracy thansmaller rain areas. STEPS models this behavior

by decomposing an observed rainfall field into arange of scales and then estimating the error in theadvection forecast as a function of scale and leadtime. Errors in the NWP also decrease as the sizeof the storms increases, so STEPS calculates theskill of the NWP as a function of scale and thenblends the advection nowcast with the NWPforecasts at each scale according to their relativeskills. The ensemble of forecasts is generated byperturbing the blended forecast by means of amodel for the forecast error in a manner thatpreserves the observed statistical characteristicsof the rain fields.

STEPS is used to generate nowcast 30member ensembles over the UK radar network, adomain of approximately 1500 km x 1000 km, at 2km, 15 minute resolution out to 6 hours andupdated hourly. The same system is used inAustralia to generate ensembles of nowcasts overa 200 km domain out to 90 minutes, updated every10 minutes but with no NWP blending andnowcasts over a 400 km domain out to 6 hours at2 km, 10 minute resolution. These ensembles areused to calculate the probability that rainfall willexceed a range of thresholds and are being testedas drivers of hydrological models so as tocalculate the probability that river levels willexceed certain critical thresholds.

3.43.43.43.4 SMS-WARRSMS-WARRSMS-WARRSMS-WARR (CMA-SMS)(CMA-SMS)(CMA-SMS)(CMA-SMS)

SMS-WARR is comprised primarily of 1) anumerical forecast model (WRF3) and 2) ananalysis/assimilation system to generate refreshedanalysis fields and to initialize the model.

The data analysis system of SMS-WARR isbased the ADAS (ARPS Data Analysis System)3-dvar package which assimilates almost allavailable data such as radar wind and reflectivity,in addition the complex cloud analysis schemefrom the LAPS was adapted into the system,

which utilizes radar and satellite observations tomake or adjust moisture field and relatedhydrometers. Together with a digital filterinitialization, initial conditions or analysesessentially make the 1-hourly cycle be awarm-start setup. The results indicate that overallspin-up time is less than 30 minutes.

The NCAR WRF ARW 3.0.1 is used asprediction model. The modeling region covers anarea about 1600 km X 1900 km with center aroundShanghai, and the resolution is 3 km by 3 km with51 levels. In terms of model physics, the cumulusparameterization was turn off and the WSM6-class micro-physics package is selected. 12-hand 3-h forecasts are made every 6-hour and1-hour, respectively, and analysis fields areprovided at each hour.

3.53.53.53.5 SWANSWANSWANSWAN (CMA)(CMA)(CMA)(CMA)

SWAN aims at being an integratedoperational severe weather nowcast platform ofCMA. SWAN ingests data from China’s newgeneration Doppler weather radars, automaticweather station, satellite, and mesoscalenumerical weather prediction model. It offers aplatform for severe weather monitoring, analysis,warning and prediction. The current version ofSWAN system provides a software package thatintegrates a series of nowcast algorithms andfunctions.

3.63.63.63.6 SWIRLSSWIRLSSWIRLSSWIRLS (HKO)(HKO)(HKO)(HKO)

The HKO nowcasting system SWIRLS hasbeen in operation since 1999. Thesecond-generation version (referred to asSWIRLS-2) has been under development andreal-time testing in Hong Kong since 2007. Aspecial version of SWIRLS-2 was implemented forBeijing and run during the WMO/WWRP B08FDPoperation period (July to September 2008) tosupport the Beijing Olympics.

SWIRLS-2 comprises a set of nowcastingsub-systems, responsible for the ingestion ofobservation data, execution of nowcastingalgorithms, as well as product generation andvisualization.

4.4.4.4. USERSUSERSUSERSUSERS

4.14.14.14.1 Forecasters:Forecasters:Forecasters:Forecasters:

Fig 4 SMS’s forecasters in the Expo operation center

The major primary users of the WENSSystems Products were the weather forecasters ofthe host city’s Meteorological Bureau – SMS. SMShas a 24-hour shift for very short range forecasting/ nowcasting. The SMS forecasters areresponsible for monitoring weather, analyzingmesoscale weather conditions, and the issue ofsevere weather nowcasts and warnings. WENSsystems provided a range of operational productsto the SMS forecasters. These operationalnowcast products for professional forecastersinclude some primary products, e.g. radar

reflectivity nowcast and convective index forecast,and some warning-related products, e.g.,quantitative precipitation forecast, probabilityforecast of the threat of severe weather, etc. Theprimary users also include the forecasters of theneighboring weather forecast offices, and theforecasters in Shanghai’s two international airportsand the Yangshan Deep-sea harbor.

4.24.24.24.2 WorldWorldWorldWorld EXPOEXPOEXPOEXPO 2010201020102010 organizationorganizationorganizationorganization committeecommitteecommitteecommitteeandandandand itsitsitsits departmentsdepartmentsdepartmentsdepartments andandandand participants:participants:participants:participants:

Final nowcasts and warnings for severeweather were directly disseminated to the WorldEXPO 2010 organizers by the SMS forecasters onduty. They organized and adjust the EXPOactivities according to SMS’s nowcasts andwarnings as necessary.

4.34.34.34.3 RelevantRelevantRelevantRelevant governmentgovernmentgovernmentgovernment departments:departments:departments:departments:

Some relevant government departments,including the Shanghai Water Bureau (SWB), theShanghai Emergency Response Centre (SERC),and some key service units, are also key users ofWENS. These second category end-users need tohave some meteorological background or to betrained in the use of WENS products. Theoperational nowcasting products for category twoindirect end users provided by WENS aregenerally regular operational nowcast products,such as those for rainfall or severe weather, andsome primary products, according to the users’needs. On receiving a warning of heavy rain andflooding, the Water Bureau would instigate urbanflood risk management. If severe weatherwarnings were issued, the Emergency ResponseCentre would start relevant preparednessactivities.

4.44.44.44.4 SpecialisedSpecialisedSpecialisedSpecialised users:users:users:users:

Some specialized users, such as media,transport and energy sectors, some private

companies, e.g. the Shanghai Yatong Co. Ltdwhich provides river transportation services inShanghai, also used the weather forecasts asguidance for their operations.

4.54.54.54.5 TheTheTheThe public:public:public:public:

Some public or private agencies orcompanies, and those organizations or agenciesthat just want to use the final nowcasting products.The products would inform the users of theimmanent occurrence of severe or hazardousweather, and what its impact might be. CategoryFour end users typically don’t have ameteorological background. They receive severeweather warnings via a range of means, such ascell phone message, TV, radio, internet, and thepublic electronic screens in streets, buildings,buses, metros, schools, and social communities.

4.64.64.64.6 UsersUsersUsersUsers andandandand theirtheirtheirtheir WENSWENSWENSWENS productsproductsproductsproducts

The WENS systems and products are directlyused by forecasters in SMS, the Expo weatherstation, and the Expo operation center. Productsare directly used by SMS’s service officers andassistants. SMS nowcasting products areindirectly used by most of the users (governmentdepartments, Expo operation center, specializedusers, and the public), while some key users canvisit the WENS page on the SMS soweather.comand directly use the nowcasting products.

5555.... FORMALFORMALFORMALFORMAL WENSWENSWENSWENS OPERATIONOPERATIONOPERATIONOPERATION

The operational phase of WENS commencedon 26 April 2010 and ran for 184 consecutive days,with nowcasting information generated by theparticipating systems providing advice andwarnings to the Expo organizers and other users.During the period of the Expo, 30 warnings ofsevere weather were issued, encompassingthunderstorms, strong winds and heavy rain.

120 120.5 121 121.5 122 122.5

Longitude(E)

30

30.5

31

31.5

32

32.5

Latit

ude(

N)

ADTD Flash DensityADTD Flash DensityADTD Flash DensityADTD Flash Density(0.2 x 0.2deg)grid(0.2 x 0.2deg)grid(0.2 x 0.2deg)grid(0.2 x 0.2deg)grid

020406080100120140160180200220240260280300320340360380400

ADTD lightning density during the Expo period (May to Oct. 2010)

Fig 5 Lightning density during the Expo period (May- Oct. 2010)

During the Expo, there were 1,709 specialweather forecasts prepared, as well as 73 specialweather warnings, of which 30 related to severethunderstorm with some high impact weather, e.g.,lightning, heavy rain, and strong winds.

After 4-month preparation on fine-tuning ofsystems and development of services andproducts, on Apr. 26, 2010，WENS stepped into fulloperation. During the 184 days, the WENSsystems were running operationally and generatedroutine products for the Expo weather services.The WENS systems and products supported theSMS forecasters and service officials as well asthe end users, including the Expo organizers andparticipants, relevant government departments(especially the emergency response agencies),some special users (particularly those in thetransport and energy sectors), and the generalpublic (including visitors to the World EXPO). Insome important Expo activities, i.e. the Expoopening ceremony, the Expo site openingceremony, the WMO Mete-Pavilion Day, etc.WENS systems performed well in monitoring,analyzing, and nowcasting. Using the WENSproducts, totally 12 lightning, 6 strong wind, and 11heavy rain warning signals were issued for the

Expo site.

Fig 6 WENS team in Expo

The Expo brought together 246 countries,regions and international exhibitors, and wasvisited by a total of more than 70 million peopleover the six months. The day of highestattendance saw just over 1 million people visit thesite. The emphasis of the Project was to helpproduce 0-6 hr forecasts and warnings ofhigh-impact weather.

The lead time of early warnings of severeweather issued was 85 minutes on average. Theoverall Probability of Detection (POD) was 83%with a False Alarm Ratio (FAR) of 17%. TheWENS proved to be good at forecastingMesoscale Convective Systems (MCS), but hadmore difficulty with local severe thunderstorms andwith tropical cyclones (although there was onlyone cyclone during the operational phase, and thispassed by at some distance from Shanghai).

Reviewing the six participating systems, andnoting that Quantified Precipitation Forecast (QPF)is a product common to all the six systems.Evaluation of the systems was not straight-forwardand was done subjectively by forecasters.Real-time verification system may be introduced tonowcasting operations in order to support

choosing the best QPF product and generating aperformance-weighted QPF product based on theperformance of each system. There was a clearneed for more impact evaluation, and improveddissemination of products and communication withusers.

An important aspect was that forecasterswould be trained to identify the strengths andweaknesses of each system so that they would bein a position to select the best product for issuing awarning in an operational context, given that thelead time in nowcasting is very short. Anotheroutcome was the need for a reliable automaticverification system. The verification schemedeveloped in CMA was a 3-year project, but it wasclear that verification was very “patchy”; it wasvery difficult to predict the accuracy of nowcastingtechniques in the short term, given that they weredealing with low-probability high-impact events.

Three severe thunderstorm cases, oneMesoscale Convective System (MCS) on July 4,one local severe thunderstorm with heavy rain andmedium hail on August 25, and one heavyrainstorm associated with tropical cyclone onSeptember 1, during the Shanghai 2010 Expoperiod are compared using the total lightning data,the WSR-88D Doppler radar data, rain gauge data,and sounding data. Time series of storm structuralparameters (e.g. vertical reflectivity core, verticalintegrated liquid products above several key levelsat 0℃, -10℃, and -20℃, and echo top) derivedfrom radar data, total lightning flash rate, cloudlightning flash rate, and cloud to ground lightningflash rate from the lightning detection data, andrain fall rate over some specific sites arecompared. The results show that some structuralinformation of storm (e.g. VILs above differentlevel, echo top, and vertical reflectivity core) andtotal lightning can be used to describe thedevelopment, intensification, and decay phases of

a storm life cycle. Very heavy rain rate is oftenfound in tropical cyclone-related storms with awarm rain process. An algorithm for identifyingprecipitation type and rain rate is being developedbased on these comparisons.

Fig 7 Time series of reflectivity (filled contours, in dBZ), height

of VILs (vertically integrated liquid) at different levels, and total

Cloud-Ground lightning (CG) and positive CG (PCG) flash rates

on Aug. 25, 2010. The bottom axis represents interval of 6

minutes of time

Case study ：On July 4, 2010, a MesoscaleConvective System (MCS) brought heavy rain,lightning strikes, and strong winds in the Expo site.In the morning of that day, severe weather earlywarning was sent to some key users, e.g., localgovernment, emergency response agency, floodprevention department, and the Expo operationcenter at 00:30 UTC, and this early warning wasupdated at 05:35 UTC. When some WENSproducts indicating high probability of severeweather in the Expo site, three types of severeweather warning signals, lightning, wind gust,heavy rain warning, were issued for the Expo siteat 05:50, 08:00, 08:40 UTC, respectively. Jointaction and cooperation among differentgovernment departments were triggered by SMSsevere weather warning. Warning information with

guidance was also sent to the public via TV, radio,web site (www.soweather.com), cell phone textmessage.

Fig 8 Reflectivity (upper) and storm cell identification and

nowcasting (lower) on July 4, 2010

6666.... WENSWENSWENSWENS EXPERIENCESEXPERIENCESEXPERIENCESEXPERIENCES

6.16.16.16.1 ExperiencesExperiencesExperiencesExperiences fromfromfromfrom thethethethe systemssystemssystemssystems

About the technical assessment of theparticipating systems and experiences learned,the WENS participating systems shared theirexperiences in Expo:

BJANC (CMA-BMS): Boundary layer

convergence is very important for stormdevelopment. Proper training of forecasters innowcasting techniques is crucial. Localized stormsposed the most difficult forecasting problems.

NoCAWS (CMA-SMS): It proved difficult toget an accurate estimation of location – this wasmuch more difficult than purely spatial resolution.There were five (5) different types of Z/Rrelationship automatically used in system, and itwas important to have a choice to best fit theactual situation.

STEPS (BOM): STEPS provides a 60 minuteaccumulation – this is really an ensemble meanrather than a deterministic forecast and gives thelowest RMS error but does not capture theextremes. There is a need to teach forecastershow to use this as it is not obvious of how tointerpret it correctly. Future changes includemoving from a server to product protocol. There isa whole system connected with STEPS forarchiving interesting data for case studies andtraining – this is a non-trivial task.

SMS-WARR (CMA-SMS): This was the onlyNWP model employed in the WENS Project. It ranon a 3 km horizontal grid with 51 vertical levels;and, good results were recorded from the modelfor an extreme temperature case (> 40 deg C) on13 August 2010. It also produced a good 9 hrforecast for dense fog. However, it was poor inpredicting a convective storm on 8 August 2010.The model was run from a warm start every hour,with a cold start at 2 p.m. (local time) each day.This cold start time was later changed to 2 a.m. toavoid starting from cold when convection wasalready well-underway. During the WENS projectthere were changes to the data assimilation. Themodel produced a short-range ensemble. However,very extreme rain was not well-forecasted by usingthe Ensemble Mean, although use of the Meanproved to be better for “normal” rain.

SWAN (CMA): SWAN under-estimatedrainfall in the Tropical Cyclone case. Usingreal-time rain gauge data for calibration would bean improvement. There are also needs to improveblending between extrapolation and meso-scaleNWP. SWAN uses different Z/R relationships indifferent installations throughout China. SWANdoes not have the capability to archive interestingsituations for case studies and training.

SWIRLS (HKO): SWIRLS used a suite ofradar products generated from the SWAN radarmosaic data by SMS – a successful collaborationexperience among the developers involved.SWIRLS produced a rich assortment of productsincluding echo-motion vectors, QPE, QPF,Probability of Precipitation (PoP), forecastreflectivity, storm-cell and severe weathernowcasts. A good appraisal was received fromSMS forecasters. Some products, in particularsevere weather nowcasts, are highly sensitive toradar data quality. The issues noted initially weremuch improved after face-to-face interactions withthe SWAN developers at the start of the WENSoperation. SWIRLS provided QPF products basedon radar extrapolation only.

Among the points which emerged from theexperience of delivering the Expo WeatherServices were: traditional forecast techniques didnot meet the unique requirements of user groups;automation as standardization needs to beenhanced; capacity-building is needed, both intechnology and human resources; themanagement of Met Services needs to bestrengthened.

6.26.26.26.2 ScientificScientificScientificScientific issuesissuesissuesissues• Initiation and evolution of severe

thunderstorms

• Relationships between a weather phenomenaand its synoptic, meso-scale, and storm-scale

conditions

• Understanding of physical and dynamicalprocesses on the micro- and mesoscales

– heavy rain in warm cloud precipitation

– thunderstorm electrification mechanisms

6.36.36.36.3 ServiceServiceServiceService andandandand EvaluationEvaluationEvaluationEvaluation• What nowcasting operation needs from

nowcasting systems and products: directconclusion, or support to user for decisionmaking;

• How to improve nowcasting service: toprovide what we can provide, or whatusers need; to do more research ondifferent requirements from users;

• How to improve the service products:nowcast text automated generation,graphic products, or new disseminationmethods.

6.46.46.46.4 LessonsLessonsLessonsLessons learnedlearnedlearnedlearned fromfromfromfromWENSWENSWENSWENS• Information delay: nowcasting based on

observation, but severe weather changesrapidly;

• How to use some new nowcastingproducts: PoP (STEPS/SWIRLS), stationPoP (STEPS), severe weather nowcasting(BJ-ANC, SWIRLS, and NoCAWS), andstorm evolution (BJ-ANC);

• Which one is the best: six nowcastingsystems with at least 5 products each atthe same time, which are not always goodor bad, e.g., QPF;

• Lack of ‘direct’ nowcasting products forwind gusts, hail, lightning, etc;

• How to improve nowcasting technology:data quality control, weather type-basednowcasting methods.

7.7.7.7. IMPACTIMPACTIMPACTIMPACT ASSESSMENTASSESSMENTASSESSMENTASSESSMENT

7.17.17.17.1 BenefitsBenefitsBenefitsBenefits fromfromfromfromWENSWENSWENSWENS

The WENS project has

• improved the Expo weather nowcastingservices, especially for the severeconvection-related high impact weather;

• improved SMS’s nowcasting operations,including upgrading the observation network,and providing new nowcasting techniques,systems, and products;

• provided very successful capacity building:forecasters got well-trained and gainedexperiences in using new systems andproviding nowcasting services;

• established a platform to exchangetechniques and experiences among theWENS participant systems;

• obtained new understanding and knowledgeof the nowcasting weather service by theactive interaction between the participatingsystems and the local forecasters;

• brought a lot of new operational proceduresbased on the idea of ‘focus on services’:

� severe weather early warning informationto some key end users was practiced inExpo;

� the feedback of these end users ispositive and praiseful;

� new communication and cooperationrelationships between SMS and endusers have been established.

• obtained new experiences and understandingof the relationship between nowcastingoperation and service;

• developed some techniques for weather

impact and service evaluation assessment;

• shared nowcasting technology and service forthe public education.

7.27.27.27.2 SocialSocialSocialSocial andandandand EconomicEconomicEconomicEconomic AssessmentAssessmentAssessmentAssessment

This was a very important aspect of theWENS Project; perhaps the first time that a full andproper social and economic assessment had beencarried out in conjunction with the scientificaspects of such a project.

The assessments were carried out throughexpert workshops, on-line surveys andface-to-face interviews. In the GovernmentAgencies and weather sensitive sectors ofassessments, the results showed that:

• The total evaluation from each user groupscored higher than 80 points for all categories.The highest scores are for satisfaction. Leadtime scored higher than accuracy;

• Type of weather is one of the factors affectingthe user evaluation; and,

• Expectation is one of the factors affecting theuser evaluation: the higher the level of userexpectation, the lower the level of satisfaction.

As for the public satisfaction of assessments,it emerged that forecast accuracy and lead-timewere deemed to be satisfactory, but that theunderstanding of forecasts was a problem.Analysis of response by level of educationrevealed that those with lower levels of educationtended to receive their weather informationthrough television, while those with higher levelsmade more use of SMS and web. In terms ofsatisfaction ratings, there was no evidentrelationship between levels of satisfaction and age,gender or level of education.

As regards to lead-time, short lead-timeswere desirable, with lead-times as low as 12

minutes fully acceptable.

Statistical analysis of the questionnairesindicated that the validity of the assessmentexercise was high. The variables that bore on“satisfaction” were: usefulness, accuracy,lead-time and content. The forecast “content” hadthe greatest influence. Understanding of theforecast was clearly very important, indicating thatthe education of the public with somemeteorological knowledge was always beneficial.

In a case-study on a power-utility, lead-timewas judged to be more important than accuracy(though both scored highly).

Among the most important conclusions of thesocial and economic assessment were:

1) The design of questionnaires was an issue;

2) Content is a critical aspect, and is closelyrelated to comprehension;

3) Proper training for the investigators wascrucial;

4) The timing of survey with respect tohigh-impact weather events had a large bearing onthe potential result.

8888.... FUTUREFUTUREFUTUREFUTURE PLANSPLANSPLANSPLANS

A challenge and opportunity after WENS ishow to translate the experiences of WENS fromExpo into research and operational focus areas forthe future. Nowcasting is still far from mature as ascience and the perceived usefulness to theoperational forecaster still needs to be addressed.

The results of the WENS DemonstrationProject will be published in a WMO TechnicalDocument as an in-depth report on WENScovering all the technical aspects and issues of theProject.

An outcome of the WENS DemonstrationProject would be helping to build the capacities ofless-developed countries in nowcasting andsharing the experiences of WENS where thesewere relevant. As a first step, the WENS FinalReview Meeting was followed by acapacity-building workshop for Cambodia, LaoPeople’s Democratic Republic, Thailand and VietNam. Material flowing from the project could beused in the preparation of WMO training activitieswhich might be undertaken for other developingcountries in the future.

According to the Final Review Meeting ofWENS, it is suggested that a Task Team (TT) maybe established to build on the results of the WENSDemonstration Project. This TT might focus ondifferent aspects of the delivery of nowcastingservices, with a special focus on impact-basedforecasting and in the context of Mega-cities.

9999.... REFERENCESREFERENCESREFERENCESREFERENCESThe official Website of Expo 2010 Shanghai China,

http://en.expo2010.cn/The official Website of Shanghai Meteorological

Service, http://www.smb.gov.cnThe Science Steering Group (SSG) of WENS,

2009. Shanghai World EXPO 2010Nowcasting Services Demonstration Project(WENS) Project Implementation Plan, WENSdocument

Final report of the Final Review Meeting for theWorld Expo 2010 Nowcasting Services (WENS)Demonstration Project and Capacity-BuildingWorkshop, Oct. 2011, Shanghai, China

AcknowledgementsAcknowledgementsAcknowledgementsAcknowledgements

The authors thank the WMO/PWS, the WENSScientific Steering Group (SSG) and WorkingGroup (WG), and the six participating systems forproviding documents and information.