1

WWRP RDP Science Plan: Tokyo Metropolitan Area Convection Study for Extreme Weather Resilient Cities (TOMACS)

Proposers: T. Nakatani1, Y. Shoji2, R. Misumi1, K. Saito2, N. Seino2, H. Seko2, Y. Fujiyoshi3 and I. Nakamura4

Institutions 1 National Research Institute for Earth Science and Disaster Prevention 2 Meteorological Research Institute 3 Hokkaido University 4 Toyo University 1. Project Summary An unprecedented dense observation campaign and relevant modeling and societal studies have been conducted since April 2010 by the National Research Institute for Earth Science and Disaster Prevention (NIED), Meteorological Research Institute (MRI), and more than 25 national institutions and universities in Japan that target local high-impact weather (LHIW) in the Tokyo metropolitan area. The objectives of the project, the Tokyo Metropolitan Area Convection Study for Extreme Weather Resilient Cities (TOMACS), include the 1) elucidation of the mechanism of LHIW in urban areas (e.g., local torrential rain, flash flood, strong wind, lightening), 2) improvement of nowcasting and forecasting techniques of LHIW, and 3) the implementation of high resolution weather information to end-users through social experiments. This proposal describes a science plan of the project as the Research and Development Project (RDP) of the World Weather Research Programme (WWRP) of the World Meteorological Organization (WMO). For the study of the mechanism of LHIW in the domestic TOMACS, data are used from the advanced observational instruments owned by participating organizations (including X-band and C-band polarimetric radars, a Ku-band fast scanning radar, Doppler lidars, microwave radiometers, a network of Global Positioning Systems (GPS), radiosondes and unmanned aerial vehicles), which are currently deployed in the Tokyo metropolitan area in addition to the operational observation networks of the Japan Meteorological Agency (JMA) and the Ministry of Land, Infrastructure, Transport and Tourism (MILT) of Japan. The intensive operational period (IOP) of the observations was set to the summers of 2011, 2012 and 2013. All observed data are archived for studies on LHIW.

2

To improve the nowcasting/forecasting techniques of LHIW, data acquired by the observation networks are also used. The X-band radars observed rainfall with 250-m resolution at 1-minute intervals. Numerical weather predictions are conducted with cloud resolving models by assimilating the meteorological data observed by several radars, lidars, GPS and radiosondes to examine how much the dense observation data can improve LHIW forecasting for urban areas. The high resolution data and forecast information from TOMACS are provided to end-users to reduce disasters caused by LHIW. For this purpose, social experiments in near real-time have been conducted. The social experiments are categorized into four fields: rescue services, risk management, infrastructure and education. The impacts of the near real-time and forecast information were analyzed by social scientists. We decided to make some parts of TOMACS an international testbed study for deep convection by proposing a RDP project relating to the WWRP working groups of Mesoscale Weather Forecasting Research (WG-MWFR) and Nowcasting Research (WGNR). The international partners include the Bureau of Meteorology (Australia), Sao Paulo University (Brazil), Environment Canada (Canada), University of Hohenheim (Germany), Pukyong National University (Korea), University Paris-Est (France), National Center for Atmospheric Research (USA) and Colorado State University (USA). The TOMACS RDP incorporates the previously mentioned scientific objectives with the study on LHIW. The international participants will conduct nowcasting/forecasting experiments collaborating with Japanese scientists using TOMACS observation data. Mechanisms of LHIW and the urban effect on their evolution are also studied. TOMACS RDP also exchanges information with other projects that share similar academic goals. A TOMACS RDP preparatory meeting was held in October 2012 with the participation of numerous scientists from six countries.

The period of the RDP project is from July 2013 to June 2016. A financial support of approximately 200 million yen per year will be supplied by the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan, for FY 2013 and FY 2014 (April 2013 to March 2015). NIED and MRI hold annual international workshops to promote communication among the participants. 2. Background

It is recognized that large cities with populations of several million people are inherently vulnerable to severe weather, such as torrential rainfall, lightning, and tornados. An increase in the occurrence of torrential rainfall and strong typhoons, which can be caused by global warming, can cause extensive damage to large cities (Ishihara, 2013). As shown in Fig. 1, the number of days with thunderstorms has been increasing in Tokyo in recent years, and the requirement of an advanced monitoring and forecasting

3

system for extreme weather is becoming greater. As a five-year research project, the domestic TOMACS has been conducted since

April 2010 by NIED, MRI and more than 25 national institutions and universities in Japan under the "Strategic Funds for the Promotion of Science and Technology" of MEXT. This project aims to understand the processes and mechanisms of extreme weather by using dense meteorological observation networks designed for the Tokyo metropolitan area in order to develop a monitoring and forecasting system of LHIW, and to implement social experiments on extreme weather resilient cities in collaboration with related local municipalities, private companies, and residents.

Figure 2 shows the monthly number of days with thunderstorms and total rainfall amounts in Tokyo from 2001 to 2010. The maximum rainfall totals, which are in September and October, are considered to be caused by typhoons and fronts, and in contrast, the maximum number of days with thunderstorms in Tokyo occurred from July to September. The field observations conducted during the IOP of TOMACS (summer 2011 to summer 2013) are shown in Fig. 3.

Fig. 1. Number of days with thunderstorms in Tokyo. Solid line indicates a 10-year

running mean.

4

Fig. 2. Monthly number of thunderstorms (bars) and total rainfall amounts (line)

in Tokyo from 2001 to 2010.

Fig. 3. TOMACS field observations.

3. RDP Proposal

We have decided to utilize some areas of TOMACS (e.g., studies on extreme weather with dense meteorological observations, and the development of nowcasting/prediction systems) in order to conduct an international testbed study for deep convection by proposing a RDP project relating to the WWRP working groups of Mesoscale Weather Forecasting Research (WG-MWFR) and Nowcasting Research (WGNR). While taking the IOP periods of the TOMACS field observations and our successive budgetary situation into consideration, we set the RDP period to three years, from July 2013 to June 2016.

5

3.1. Observation network One of the unique features of TOMACS is the utilization of dense meteorological

instruments in the Tokyo Metropolitan area, which is one of the most urbanized areas in the world. The field campaign that was planned by 14 research organizations, which was initiated in summer 2011 and will end in summer 2013, targets the atmospheric environment of the troposphere and boundary layer, and the initiation of convections and lifecycles of thunderstorms. Details of the observations are given in the Appendix (A1-A3). 3.2 Data archive and policy TOMACS observation data are archived into data servers at MRI and NIED. Some JMA and MILT data are accessible from MRI via internet. To promote the use of such data, file conversion kits that convert the original data format to a community format (NetCDF) have been developed. International registered participants are regarded as domestic TOMACS participants and follow the same data policy as Japanese scientists. As for other special observation data, the observers keep their priority for the analysis, and such data are generally offered individually by a mutual agreement between the observers and users. Data taken by the most advanced systems are, in principal, not available. For details, see the Appendix (A4). 3.3. Nowcasting and numerical modeling Currently, one of the most challenging areas in mesoscale meteorology is the prediction of LHIW caused by mesoscale convective systems via high resolution data assimilation. Understanding the mechanisms of LHIW and the urban effect on their evolution are also important scientific areas. Observational data, nowcasting, and numerical modeling, including data assimilation to reduce the gap between nowcasting and the initial condition of numerical models, are key factors for the prediction of LHIW.

As a part of TOMACS, quantitative precipitation nowcastings (QPN) with X-band polarimetric radars are conducted by NIED. The Short Term Ensemble Prediction System (STEPS) owned by the Australian Bureau of Meteorology is also implemented as an advance nowcasting system. As for data assimilation techniques, cloud resolving variational methods (3D-VAR and 4DVAR) and the Local Ensemble Kalman Filter (LETKF) will be applied by MRI, NIED, and the international participants. For details, see the Appendix (A5 and A6). 3.4. Social experiment and liaison with the Dallas-Fort Worth Experiment

TOMACS also aims to implement social experiments on extreme weather resilient cities in collaboration with the related governmental institutions, local governments,

6

private companies, and residents. The social experiments will be carried out in four study fields: (1) rescue services, (2) risk managements, (3) infrastructure, and (4) education. The main purpose of the social experiments is to evaluate how the advanced weather information will provide effective warnings, and promote proper evacuations and rescue activities. For details, see the Appendix (A7). TOMACS exchanges information with other projects that share identical academic goals. Liaison with the Dallas-Fort Worth (DFW) Urban Test-bed will be prompted by the international participants from Colorado State University (CSU) (see Appendix A8). 3.5. Preparatory meeting in October 2012

In preparation of the RDP proposal, a TOMACS RDP preparatory meeting was held in Shinagawa Intercity, Tokyo, Japan, on 25 October 2012. Approximately 50 participants from six countries attended the meeting (Fig. 4). Details of the meeting are given in the Appendix (A9).

Fig. 4. Group photo of the TOMACS RDP preparatory meeting (25 October

2012). 4. RDP Management and organization 4.1. RDP management

The research and development project is managed by NIED and MRI, with participation from the domestic/international scientists and advisors. 4.2. International participants

Table 1 shows a list of the tentative international participants of the TOMACS RDP, which includes seven institutions from six countries. Additionally, several other scientists from the USA and Korea are currently showing interest in becoming participants.

7

Table 1. List of international TOMACS RDP participants

Country Institution Participant Area Counterpart Australia Bureau of

Meteorology (BOM) A. Seed Nowcasting NIED

Brazil Sao Paulo Univ. A. Pereira Nowcasting Urban modeling

MRI

Canada Environment Canada (EC)

S. Belair

Urban modeling MRI

Germany Univ. of Hohenheim (HU)

V. Wulfmeyer Variational data assimilation

MRI

Korea Pukyong National Univ. (PKNU)

D.-I. Lee Disdrometer observation

NIED

France UANPE D. Schertzer Precipitation Kagoshima Univ. USA

NCAR J. Sun Storm-scale data assimilation

CRIEPI

Colorado State Univ. (CSU)

V. Chandrasekar Dallas-Fort Worth testbed

Kagoshima Univ.

4.3. International Science Steering Committee

The International Science Steering Committee (ISSC) currently has a membership that consists of the part of the proposers and international participants. A tentative list is given in Table 2.

Table 2. List of the ISSC members (tentative) Country Institution Name Roles Japan NIED T. Nakatani TOMACS principal

investigator (PI) R. Misumi TOMACS management

MRI Y. Shoji Field campaign K. Saito RDP proposal to WWRP

Australia BOM A. Seed Nowcasting Brazil Sao Paulo Univ. A. Pereira Radar meteorology Canada EC S. Belair Urban modeling Germany HU V. Wulfmeyer Data assimilation Korea PKNU D.-I. Lee Field observations USA NCAR J. Sun Data assimilation

4.4. International Advisory Board

The International Advisory Board (IAB) currently has a membership that consists of the proposers, main contributors, international participants and international experts. A tentative list is given in Table 3.

8

Table 3. List of the IAB members (tentative) Country Institution Name Remarks Japan Toyo Univ. I. Nakamura Social experiment

NIED K. Iwanami Collaborator of RDP science plan Hokkaido Univ. Y. Fujiyoshi Proposer of RDP science plan Kagoshima Univ. M. Maki Former PI of TOMACS Kyoto Univ. M. Ishihara Proposer of TOMACS

Canada EC P. Joe Chairman, WGNR Netherlands KNMI J. Onvlee Chairman, WG-MWFR France UANPE D. Schertzer Precipitation USA CSU V. Chandrasekar PI, Dallas-Fort Worth project

4.5. Local Organizing Committee

The Local Organizing Committee (LOC) includes the managers for TOMACS as a WWRP RDP. A tentative list of the LOC members is given in Table 4.

Table 4. List of the LOC members (tentative)

Institution Names NIED S. Suzuki*, Y. Shusse, K. Hirano and A. Nakai MRI H. Seko*, N. Seino, M. Otsuka, M. Kunii, and N.

Imai * Head of the LOC members at NIED and MRI.

5. Milestone

The IOP of the TOMACS domestic field campaign was set as the summers of 2011-2013; however, some observations may continue until 2014. Considering the successive budgetary situation, we set the RDP period to three years, from July 2013 to June 2016. The first RDP International Workshop will be held in December 2013 in Tsukuba, Japan, and workshops are scheduled to take place each year until 2015.

Table 5. Schedule of TOMOACS

2013 2014 2015 2016

Observation

Data archive

Forecast

experiments

Social

experiments

International

meetings

Summary

9

6. Budgetary background In FY 2013 (April 2013-March 2014) and FY 2014 (April 2014-March 2015), the Ministry of Education, Culture, Sports, Science and Technology, Japan, will provide support of approximately 200 million yen per year for the domestic studies conducted by TOMACS. After April 2015, the budgets for international meetings will be sought by NIED. 7. Societal impact

Elucidation of the mechanism of LHIW in urban areas will contribute not only to the improvement of forecasting and nowcasting of severe weather, but will also provide awareness of high-impact weather through an increase in knowledge. Moreover, the forecasting and nowcasting techniques that will potentially be developed with this project will be applicable to many countries where similar disasters occur in urban areas. The circulation of such advanced weather information will contribute to the resilience of natural disasters as well as an improvement in the quality of life for people living in urban areas. Acknowledgments The proposers express their thanks to Dr. Kojun Yamashita and Dr. Shigenobu Matsuda of the Japan Science and Technology Agency for their support towards TOMACS. We are grateful to the domestic participants of TOMACS: Y. Okada, K. Iwanami, S. Suzuki, T. Maesaka, T. Kayahara, S. Shimizu, T. Wakatsuki, Y. Shusse, K. Hirano, K. Nakane, Y. Usuda and H. Taguchi of the National Research Institute for Earth Science and Disaster Prevention; K. Kusunoki, A. Adachi, H. Yamauchi, E. Sato, H. Inoue, S. Saito, Y. Yamada, S. Nagumo, F. Fujibe, H. Ishimoto, T. Kawanbata, S. Origuchi, M. Otsuka, M. Kunii and S. Yokota of the Meteorological Research Institute; M. Kawasaki, Y. Kikumori and S. Tsuchiya of the National Institute for Land and Infrastructure Management; N. Sekiya of Toyo University; A. Yamaji, H. Nakagaki, T. Momotani and A. Goto of the Japan Weather Association; M. Kawashima of Hokkaido University; M. Yasui, K. Ishii, K. Mizutani, H. Iwai, H. Hanado, S. Kawamura and S. Sato of the National Institute of Information and Communications Technology; Y. Suzuki and M. Homma of Kyoto University; T. Ushio, S. Yoshida and Z. Kawasaki of Osaka University; H. Nakamori and M. Fukuda of Nihon University; T. Hoshikawa of Otsuma Women's University, T. Sano, K. Soma and K. Sunada of the University of Yamanashi; T. Yamada and K. Yoshimi of Chuo University; F. Kobayashi and H. Sugawara of the National Defense Academy; R. Oda of Chiba Institute of Technology; T. Yoshihara of

10

Electronic Navigation Research Institute; H. Hirakuchi, S. Sugimoto, Y. Toyoda, M. Nomura and A. Hashimoto of the Central Research Institute of Electric Power Industry; H. Yokoyama, A. Ichihashi, T. Fujiwara, H. Ando, Y. Seto and K. Hiroi of the Tokyo Metropolitan Research Institute for Environmental Protection; N. Ehara, K. Mizumura, H. Nakano, M. Yoshii, A. Iso and N. Takahashi of the Tokyo Fire Department; K. Takai, N. Tachihara, K. Hasegawa and R. Itabashi of the Edogawa City Government; T. Kirihara, H. Chiba, E. Okamoto, N. Osanai, K. Inoue, K. Sato and Y. Tsuda of the Yokohama City Government; K. Shiobara of the Fujisawa City Government; M. Hoda and Y. Suya of the Minami Ashigara City Government; A. Togari of the East Japan Railway Company; D. Tsujii of the Central Japan Railway Company; K. Ohtsuka of Obayashi Corporation; T. Murano, F. Mizutani, H. Ishizawa, T. Kashiwagi, M. Wada and T. Mizuno of the Toshiba Corporation Social Infrastructure Systems Company; M. Yogo and H. Onishi of the Certified and Accredited Meteorologists of Japan. We also give thanks to the Water and Disaster Management Bureau of the Ministry of Land, Infrastructure, Transport and Tourism, and the Observation Department and Meteorological Satellite Center of the Japan Meteorological Agency for their understating and cooperation with this project in regards of the observational data in this study.

1

Appendix Supporting Documentation A1.1. Field observations

One of the unique features of TOMACS is the utilization of dense meteorological instruments in the Tokyo Metropolitan area, which is one of the most urbanized areas in the world. The field campaign that is currently underway in the Tokyo metropolitan area uses research instruments and operational meteorological networks, and was planned by 14 research organizations. The field observations began in the summer of 2011 and will end in the summer of 2013. Their objectives are to target the tropospheric environment, boundary layer, initiation of convections and the lifecycles of thunderstorms. Observations of the convective environmental conditions are carried out via radiosondes, wind profilers, a GPS network, an unmanned air vehicle, and a network of automated weather stations. The generation and development of convective precipitation are investigated by evaluating observations conducted by a Doppler lidar, high repetition scanning geostationary satellite, Ku-band polarimetric radar, X-band polarimetric radar network (X-NET), C-band research polarimetric radar, and C-band operational Doppler radars.

The dense meteorological observations in the Tokyo metropolitan area are currently being considered as an international test-bed for a study on deep convection. A list of the observation instruments and their targets is shown in Table A1.

Table A1. List of observation instruments and their targets.

Targets Parameters Instruments

Environment T, wind, water vapor in the Troposphere

Research aircraft Radiosonde stations

GPS receivers

UHF wind profilers

Micorwave radiometers

Boundary layer

Air flow Doppler Lidars

Vertical heat flux Scintillometer Radiation measurement

Surface T, wind, rain (partial Td, P)

High-resolution surface network

JMA AMeDAS

River Bureau rain gauges

Cumulus clouds

Images of cumulus clouds

MTSAT rapid scan imager

Web cameras

Thunderstorms Fast-scan, high-resolution 3D reflectivity/velocity

Ku-band fast scan radar

3D reflectivity/velocity

X-NET MP radars

JMA C-band Doppler radars

MRI C-band Polarimetric radar

2

A1.2. Observation field

The project test-bed consists of the Tokyo metropolitan area (Fig. A1), which is defined in the present study as a 50 km area surrounding the headquarters of the Tokyo Metropolitan Government. The Tokyo metropolitan area is the world’s most populous metropolitan area: the total population is approximately 30 million, which corresponds to the entire population of Canada. About 10 million people use public transportation every morning to commute to their schools and/or offices. The area comprises most of Tokyo, Kanagawa, and Saitama Prefectures and a part of Chiba Prefecture. Five mega-cities are within this area: Tokyo (8.5 million), Yokohama City (3.6 million), Kawasaki City (1.3 million), Saitama City (1.2 million), and Chiba City (0.9 million). The network of radars in the Tokyo metropolitan area is one of the unique features of the present study.

Fig. A1.1. Observation field of TOMACS. Numbers show population in millions.

A2.1. Operational observations by JMA and MILT

JMA has nationwide upper-air sounding networks. Operational radiosonde

Polarimetric parameter

X-Net MP radars MLIT X-band MP radars

MRI C-band Polarimetric radar

Distribution of raindrops

2D-Disdrometer 1D-Disdrometers

Lightning JMA Lightning Detection System

3

observations take place at 16 stations in Japan twice a day and UHF wind profilers at 31 stations are operated within the WINDAS system to measure wind profiles every 10 minutes (Fig. A2.1, left panel). JMA's network of 20 C-band radars (each with a wavelength of 5.6 cm) covers a majority of Japan and observes the intensity and distribution of rainfall (Fig. A2.1, right panel). Radar data are digitized to produce special radar-echo composite maps every 5 minutes for the purpose of monitoring precipitation throughout the country. The data are also calibrated with AMeDAS in-situ data and rain gauge data from related authorities for use as initial values in very short range forecasting of precipitation. JMA has been replacing its radars with Doppler units that can measure not only precipitation intensity but also the radial velocity of raindrops. As of the end of March 2010, 16 out of the 20 units in the network had been replaced with Doppler radars.

Surface observations are carried out at about 1,300 stations via automatic observation equipment, which are collectively known as the Automated Meteorological Data Acquisition System (AMeDAS). Stations are laid out at average intervals of 17 km throughout the country, with about 1,200 of them unmanned. Observations at manned stations cover weather, wind direction/speed, amount of precipitation, type and base height of clouds, visibility, air temperature, humidity and atmospheric pressure. Data other than those relating to weather, visibility and cloud-related elements are automatically sent every ten seconds from manned stations and some unmanned stations, and every ten minutes from the remaining unmanned stations, of which about 700 observe four elements (precipitation, air temperature, wind direction/speed and sunshine duration) and about 300 observe precipitation.

Fig. A2.1. (Left) Radiosonde stations (green squares) and wind profilers (red squares) of JMA. (Right) Radar observations conducted by JMA. Red circles indicate Doppler radars.

4

Fig. A2.2. JMA’s AMeDAS network. A.2.2. MTSAT

Rapid scans, which are performed every 5 minutes, from JMA’s geostationary satellite MTSAT are used to provide useful information on the initiation of thunderstorms. The convective cloud information observed by MTSAT and the information on precipitable water distribution given by GEONET will be used to study the initiation of convections and the ensuing processes that develop. Test operations of MTSAT (Himawari) rapid scans began in the summer of 2010. Since the summer of 2011, visible and IR images have been provided to the experiment every 5 minutes. The rapid scan images are expected to elucidate rapid changes in the initiation of convection.

Fig. A2.3. Rapid scanning area of MTSAT.

5

A.2.3 XRAIN (X-band polarimetric Radar Information Network) Torrential rainfalls cause water levels to rise rapidly and result in inundation damages, especially damages caused by rivers in urban areas. Thus, there is a need for a more detailed and rapid distribution of information that can be used for awareness in terms of evacuations and disaster prevention. In light of such situations, X-band polarimetric RAdar Information Network (XRAIN) is currently being developed. As of September 2012, 27 radars are in the operational test phase, and eight more will begin their operational tests in 2013. The X-band radars enable observations with smaller grid sizes when compared to the C-band radars. Furthermore, these radars utilize two types of radio waves: vertical and horizontal polarization. This allows for an improvement in accuracy by allowing raindrops to be characterized by their shapes, which change from round to flat as raindrop size increases. This eliminates the need of ground-based raingauge data for data correction, and enables a more rapid dissemination of information.

Fig. A2.4. XRAIN status (as of September 2012). A3.1. Observations by TOMACS

The following research radars and operational radars are located in the Tokyo metropolitan area: X-NET (five X-band polarimetric radars and three Doppler radars), two MLIT (Ministry of Land, Infrastructure, Transport and Tourism) X-band polarimetric radars, an MRI C-band polarimetric radar and three JMA C-band operational Doppler radars. a. X-NET

X-NET is an advanced X-band research radar network that has been implemented by the National Research Institute for Earth Science and Disaster Prevention and several universities and research institutes to mitigate urban disasters caused by severe storms. Figure A3.1 shows the locations of the research radars that make up X-NET; the network

6

consists of three NIED polarimetric radars, a Chuo University Doppler radar, an NDA Doppler radar, and a JWA Doppler radar. All the research radars have the X-band wavelength. The EBN radar in Kanagawa Prefecture has been in operation since 2003. All data from the connected radars are sent to NIED in Tsukuba via an optical communications system. The NIED MKA polarimetric radar, which was installed between June to October 2009, is transportable. The CRIEPI and University of Yamanashi X-band polarimetric radars were incorporated into X-NET in 2011. The X-NET observation area is approximately 200×200 km. The basic antenna scan mode is volume scan, with a 5-minute repetition time. The horizontal resolution, which is dependent on the characteristics of the individual radar system, has a range of 100-200 m and an azimuth angle of about 1.3°.

Fig. A3.1. X-NET radars and observation area

X-NET data are classified into six levels, ranging from 0 to 4, in accordance to their level of processing (for details, see section A4.1). b. Doppler Lidars

Urban climate is very complex due to the effect of flow distortions caused by buildings and other anthropogenic structures. The characteristics of urban surfaces at the bottom of the atmospheric boundary layer have been studied intensively in the past few decades. This is because the physical processes in the urban surface layer directly influence the atmosphere above, and thus understanding and modeling of these physical processes are crucial for progress in urban meteorology. An understanding of turbulence is crucial for

7

directing future developments in urban surface modeling. Doppler lidar is a relatively new addition to the technology of atmospheric wind sensing. One of its advantages is the ability to use aerosol particles as atmospheric scattering targets. Due to the negligibility in the fall speed of aerosols, they are excellent tracers of air motion. A 3D-Coherent Doppler Lidar (3D-CDL) is the only device that can seamlessly measure 3D distributions of aerosols, wind and clouds. It is a powerful tool to study not only the structure of the boundary layer, but also the formation mechanisms of low-, mid-, and upper-level clouds.

Two research Doppler lidars (NICT and Hokkaido University), as well as the JMA operational Doppler lidars at the Narita and Haneda airports (two lidars at each airport), are operated to observe the behavior of sea breeze fronts and air flow in the atmospheric boundary layer prior to the initiation of convection. A 3D-CDL operated by Hokkaido University uses a wavelength of 1.54 μm (eye-safe) (Mitsubishi Electric Corporation).

Fig. A3.2. (Left) Doppler lidar at Hokkaido University. (Right) An example of Doppler

lidar PPI (EL: -0.5 deg.) scanning patterns for 9 August 2011 from 8:14-8:55 JST (Day-time).

c. Microwave radiometers

Microwave radiometers are currently being operated by Hokkaido University and MRI to observe the vertical profiles of temperature and water vapor, integrated water vapor and liquid water path. The RPG-HATPRO supports two temperature profiling modes: full tropospheric profiling (0-10 km) and boundary layer scanning (0- 2 km). Humidity profiling is only available with the full tropospheric mode. The vertical resolution of the retrieval outputs for any boundary layer temperature profile is 50 m, and the profile accuracy is +/– 0.7 K RMS. Data are obtained every 5 minutes.

Radiometer observations are also conducted by MRI.

8

Fig. A3.3. (Left) Microwave radiometer (RPG-HATPRO) at Hokkaido University. (Right) An example of a boundary layer temperature profile from 20 July 2011 to 10 August 2011.

d. GPS/GNSS observations

The global navigation satellite system (GNSS), which utilizes the U.S. Global Positioning System (GPS), is fully adapted for the experiment to obtain information on atmospheric water vapor, which is a primary source of convection. The Geospatial Information Authority of Japan (GSI) operates the nationwide ground-based GPS earth observation network (GEONET) with more than 1,200 GNSS stations. JMA has controlled the retrieval and assimilation of precipitable water vapor (PWV) from GEONET since October 2009. Since July 2011, five receiver sets have been installed in the eastern region of the Tokyo metropolitan area to increase the spatial resolution of the GNSS in the TOMACS experiment. One of the stations is located on an artificial island called “Umihotaru” (translated as “Sea Firefly”) in Tokyo Bay. The sampling rate and elevation cutoff angle are set to 1 s and 0 degrees, respectively. Additionally, two GNSS stations in Tokyo can be utilized that are registered to the global tracking network of the International GNSS Service. Furthermore, the Earthquake Research Institute at the University of Tokyo kindly offers GNSS data that are observed on the institute’s rooftop.

The GIPSY-OASIS II (Webb and Zumberge 1993) software package is used to analyze the zenith total delay (ZTD), atmospheric gradient parameter, and the slant total delay (STD) every 5 minutes at the above mentioned sites. PWV time series are converted from that of ZTD by using the temperature and pressure at each GNSS station as interpolated from JMA’s surface meteorological observation network. In addition to the traditional GNSS atmospheric parameters, two new indices that describe the spatial concentration of water vapor and the higher-order water vapor inhomogeneity are proposed, and their relationship with convective precipitation is assessed (Shoji 2013).

9

Fig. A3.4. GPS/GNSS stations in the Tokyo metropolitan area.

e. Radiosondes

Radiosonde soundings provide aerological data in higher temporal resolution for the TOMACS IOP at several sites in Tokyo and its surrounding areas (Fig. A3.5). The radiosonde soundings hold two objectives. One is to capture the atmospheric environment during heavy rainfall. The other is to investigate the structure of the urban boundary layer and its role in the formation of convective systems. Because the observations were planned to be conducted in a densely populated area, the sonde’s landing area is predicted via a sonde tracking simulation before each launch. A severe thunderstorm hit Tokyo during the afternoon of 26 August 2011. A maximum

hourly precipitation of more than 80 mm/h was observed at Haneda airport and its neighboring stations. Radiosonde observations took place at Tsukuba and Yokosuka and data of the atmospheric conditions for this extreme event were obtained. An intensive observation of the urban boundary layer was conducted from 27 September to 7 October 2011 at Tsukuba, Ukima (Tokyo-E), Koganei (Tokyo-W) and Yokosuka (Fig. A3.5). Ukima is the closest site to central Tokyo. A GPS radiosonde was launched nearly every three hours from 9:00 to 21:00 JST (Japan Standard Time = UTC + 9 h). The observation results provide new insight into the characteristics of the urban boundary layer. Cooperative radiosonde observations with a JMA weather observation ship are planned for summer 2013. The planned observation area is to the south of Tokyo, the sea off the coast of Kanto area.

10

Fig. A3.5 (Left) Radiosonde observation sites and (right) major observation periods for TOMACS 2011.

Fig. A3.6. Photo of the JMA weather ship that is planned to conduct cooperative observations in summer 2013.

f. Aircraft observations At first, a Beech Craft B99 of the Electronic Navigation Research Institute was planned to conduct the TOMACS aircraft observations in order to observe the detailed environmental conditions around a cumulonimbus cloud over the Tokyo metropolitan area. However, the aircraft was destroyed by a tsunami caused by the Great East Japan Earthquake on 11 March 2011. As an alternative plan, an unmanned aerial vehicle (UAV) has been used since 2012. Images and specifications of the UAV are shown in Fig. A3.7. The UAV is a twin-engine aircraft, 4.6 m in width, 35 kg in weight, and was manufactured by Fuji Imvac Inc. Its cruising speed is 100-120 km/h with a 4-hour flight duration. The twin engines allow the aircraft to fly stably, even when one engine accidently stops. Moreover, a parachute is released during emergencies. Temperature and relative humidity are measured by a RS-06G sensor made by Meisei Electric. Wind components are

Tsukuba

Tokyo-E Tokyo-W

Yokosuka

JST 03 06 09 12 15 18 21 24 03 06 09 12 15 18 21 24 03 06 09 12 15 18 21 24TsukubaUkima

KoganeiYokosuka

JST 03 06 09 12 15 18 21 24 03 06 09 12 15 18 21 24 03 06 09 12 15 18 21 24TsukubaUkima

KoganeiYokosuka

JST 03 06 09 12 15 18 21 24Tsukuba routine observation at the Aerological Observatory

UkimaKoganei

Yokosuka

10/7（Fri）radiosonde observation from this experimen

9/27（Tue） 9/28（Wed） 9/29（Thu）

10/3（Mon） 10/4（Tue） 10/5（Wed）

11

estimated with a GPS and an inertial measurement unit. The accuracy of the estimated wind vectors will be verified by a comparison with pilot balloon observations. The observations in 2013 will be conducted in Sagami Bay, which is the source of water vapor for the Tokyo metropolitan area during the summer. Observed data will be input into a data assimilation system and their impact on forecasting thunderstorms will be evaluated.

Fig. A3.7. Images of the UAV and the measurement specifications.

g. In-situ boundary layer observations One of the keys to a successful experiment is the proper measurement of the

atmospheric boundary layer, since the formation of a convective system is affected by a variety of processes in the boundary layer. TOMACS tackles this with various types of instruments and techniques, as mentioned in this chapter. The development of the boundary layer is closely related to the thermal forcing near the land surface. In the Tokyo area, the influence of anthropogenic heating should also be considered. To evaluate the surface heat budget in an urban area, sensible heat flux measurements with a large aperture scintillometer have been conducted in the Itabashi Ward of Tokyo. The relationship between the surface heat flux and the height of the boundary layer affecting the initiation of cumulus clouds is investigated by utilizing a dataset obtained by radiosonde soundings.

12

Fig. A3.8. Schematic of the boundary layer observations.

A3.2. Advanced observations a. MRI C-Band MP radar

A MRI C-band polarimetric radar is a prototype system that was originally designed to reduce the occupied bandwidth necessary for meteorological observations. This type of radar is mounted on top of the MRI building in Tsukuba, which is located about 50 km northeast of Tokyo, Japan. This system employs two solid-state transmitter units to simultaneously emit horizontally and vertically polarized waves. The high stability of the radio waves generated by the solid-state transmitters enables the radar to make polarimetric observations with high time-resolution and reliability.

Because the peak power of the transmitter is slightly weak, observations are made with a long pulse to increase the average power. A pulse compression technique is used to increase range resolution. One of the issues associated with this technique is the generation of range side lobes. To overcome such problems, we have been studying the optimizations of operating parameters, such as transmitting waveforms and developing data processing algorithms.

13

Fig. A3.9. MRI C-Band MP radar in Tsukuba, Japan.

b. Ku-band radar

A Ku-band “fast scan” radar was developed by Osaka University and Sumitomo Electric Industries in order to obtain a 3-D structure of a cumulonimbus cloud every 1 min. The data observed by this radar are useful not only for developing a nowcasting system in narrow area, but also for the clarification of the mechanisms of severe phenomena. For example, it is thought that the descending reflectivity core (DRC) plays an important role in the generation of a tornado or a downburst. We are currently operating the radar at Seikei University in Musashino-shi, Tokyo, in order to capture heavy rainfall events in the Tokyo metropolitan area.

Fig. A3.10. The Ku-band radar in Musashino-shi, Tokyo.

c. AWS network

In addition to JMA’ s operational automated weather station (AWS) network (AMeDAS), a network of 12 AWS stations with 3 km resolution (Fig. A3.11) was set up in the western part of the TOMACS radar observation area. The observation intervals are 1 second for wind direction and wind speed, and 10 seconds for temperature, humidity, and pressure.

14

Six laser-based optical disdrometers were mounted onto six stations to measure drop size distribution and rainfall intensity during the 10-second interval. The surface data are being used to study the detailed structure of extreme weather and to validate radar QPE and QPF algorithms.

Fig. A3.11. (Upper) Dense TOMACS AWS network. (Lower) Example of an observation

site.

15

A4. Data A4.1. Data archive a. X-NET products X-NET data are classified into six levels, ranging from 0 to 4, in accordance to their level of processing.

• Level 0 includes raw radar data that are usually recorded in the original format used by each radar manufacturer.

• Level 1 data are the raw data after their format has been converted to NetCDF, which is a set of software libraries and machine-independent data formats that support the creation, access, and sharing of array-oriented scientific data.

• Level 1.5 data are the radar data after fundamental corrections have been applied, such as attenuation correction and unfolding of Doppler velocity data. The accuracy of the correction scheme for this level has a large impact on the data of higher levels, such as those for rainfall rate and wind speed. The data format used for Levels 1 through 4 is NetCDF. For Levels 1.5 to 4, geographic coordinates are used for the coordinate system, which is a more convenient system for the application of data.

• Level 2 data are essentially real-time data, and refer to the “basic” meteorological products calculated from Level 1.5 data. Level 2 data can be classified as 2D or 3D, which are denoted as “2p” and “2v”, respectively. The same rule is applied to higher level data. Examples of Level 2 data include distributions of rainfall, rain water content, wind speed, and wind direction.

• Level 3 data are “advanced” meteorological data calculated from Level 2 or Level 1.5 data, and examples of such data include distributions of the area of rainfall (AR) within a drainage area, cumulative rainfall (CR), effective rainfall (ER), and vertically integrated liquid water content (VIL). Other examples include microphysical parameters, such as the hydrometeor type (HT) and drop size distributions (DSD). Objectively analyzed wind fields adjusted with a numerical model are also categorized into the Level 3 products. Additional examples include echo top height, kinematic parameters, and wind gust data.

• Level 4 data are “forecasted” products. Graphical examples of the products of each level are shown in the next chapter of this paper. While the X-NET products are basically retrieved from research radar data, some X-NET products depend on numerical forecasting data and operational weather information. These supplemental data are also shown in Table A4.1.

16

Table A4.1. List of the X-NET products.

b. Data archive Observation data are archived at the MRI and NIED data servers. Some JMA data are accessible from MRI via JMA’s intranet. To promote the use of the data, file conversion kits that convert the original data format to NetCDF have been developed for TOMACS. Table A4.2. Data archived at the MRI server.

Data type Archive server

Original format

Resolution, Period

Comments

Rapid Scan MTSAT data

MRI server NetCDF VIS 0.01 degree IR 0.04 degree

① A1 DRAW Airport Doppler radars (Narita and Haneda)

MRI server GRIB2 500 m, 1.0 degree

NetCDF converter (radar_rt2nc) completed for (r, θ) or (lon, lat) coordinates

① A2 JMA C-band Doppler radar (Tokyo)

MRI server GRIB2 500 m, 1.0 degree

NetCDF converter (radar_rt2nc) completed for (r, θ) or (lon, lat) coordinates

② Airport Lidars (Narita, Haneda)

JMA DVD-R (to be archived at MRI server)

SIGMET (convertible to DRAFT）

NetCDF converter (lidar2nc) prepared by modifying the converter for DR

MLIT MP radars (Saitama, Yokohama, Shizuoka and Niigata)

MRI server Original NerCDF converter completed

17

Table A4.3. X-NET data archived at NIED.

Data type Archive server Format Comments Level 0

NIED server NetCDF Raw radar data recorded in the original format of a radar company.

Level 1

Raw data after the format conversion to NetCDF.

Level 1.5

Radar data after fundamental corrections, such as attenuation correction and unfolding of Doppler velocity data.

Level 2 “Basic” product such as rain rate and wind speed.

Level 3

“Advanced” data such as VIL, areal rainfall, etc.

Level 4 “Predicted” data of precipitation and wind. Table A4.4. JMA data accessible at MRI.

Data type Archive server Format Period Comments ⑦AMeDAS JMA server Original

(binary) 10 min NetCDF converter

(amd2nc) prepared with reference to UNIDATA format

⑤⑥ Surface JMA server 1 min or 10 min NETCDF converter (sfc2nc, sfctm2nc) was used

④ A1 Sonde data from JMA Aerological Observatory

2 sec (?) NETCDF converter (sonde2nc) prepared by modifying the converter for WPR

③JMA wind profilers (e.g., Mito)

JMA server 10 min NETCDF converter (wpc2nc) was used

LIDEN Original format (text)

GPS total precipitable water

hourly Processed by JMA

Table A4.5. NWP data archived at MRI.

Data kind Archive server Format Resolution, Period

Comments

⑧ JMA mesoscale analysis

MRI server NuSDAS 5 km, 3 hourly (pressure-plain, land, model plain)

NetCDF converter (nusdas2nc) was used

⑨JMA global model forecast (mfboundary)

MRI server NuSDAS 5 km L50, 1 hourly

NetCDF converter (nusdas2nc) was used

⑩JMA hourly analysis

JMA server (accessible from MRI)

NuSDAS 5 km L50 1 hourly

NetCDF converter (nusdas2nc) was used

18

Table A4.6. Other provided data. Data type Archive

server Format Resolution,

Period Comments

GPS total precipitable water

MRI PC (Shoji)

Original (text)

5 min Processed by Shoji

④ A2 Sonde data from TOMACS

MRI PC, each observation site

Original (binary)

2 sec NerCDF converter (sonde2nc) was used

NICT Doppler lidar

NetCDF

Hokkaido Univ. Doppler lidar

Original (convertible to DRAFT）

about 1 min NETCDF converter (lidar_rt2nc_h) was used

Table A4.7. Other data not provided.

Data type Archive server Format Period Comments MRI C-band radar

MRI PC Original about 4 min

MRI Ku-band radar

Original (convertible to DRAFT）

about 1 min

Meso network MRI PC (Kusunoki)

A4.2. Data policy

In TOMACS, international registered participants are regarded as the participants of the “Social System Reformation for Adaptation to Climate Change,” Strategic Funds for the promotion of Science and Technology (JST/MEXT), and follow the same data policy as the domestic participants. An overview of the data policy is as follows.

• Quick-view data can be opened on-line (reproduction is prohibited).

• X-NET data will become available after 2 years of observations. • JMA and MLIT data are available (agreement required). • As for other special observation data, the observers hold all rights priority for use.

This type of data is basically offered individually by a mutual agreement between the observer(s) and user(s).

• Data taken by the most advanced systems described in section A3.2 (i.e., the C-band polarimetric radar, KU-band radar, and the dense AWS network) are, in principal, not available to international participants.

A5. Nowcasting

The high resolution X-band polarimetric radars (250-m mesh and update at 1-minute interval) that cover the entire Tokyo metropolitan area are expected to improve the quantitative precipitation nowcasts (QPN). In TOMACS, the following three techniques

19

will be applied. The first is an extrapolation technique. This technique is robust and classic, and conducting QPNs with X-band polarimetric radars will show better results than that with C-band conventional radars. NIED, JWA and Toyo University are planning to conduct QPNs based on the extrapolation technique and deliver the data to people via cell phone text messages. The second technique utilizes vertically integrated liquid-water content (VIL). It is confirmed that VIL derived from the specific differential phase (KDP) of the polarimetric radars shows a 5-10 minute earlier change in the values than the ground-based rain gauges (Hirano and Maki 2010). Such characteristics would be useful when making such a forecast when providing warnings of a downpour to people working or playing outside. An algorithm that detects precipitation cores is also under development. The third application is the Short Term Ensemble Prediction System (STEPS) owned by the Australian Bureau of Meteorology. It is considered to provide very high resolution forecasts that use low resolution radar data to extend the lead times of the forecasts. Furthermore, social experiments on how to make decisions when using ensemble forecasts are also planned.

A6. Numerical modeling A6.1. Urban modeling

Numerical modeling is the key to understanding the mechanism of extreme weather. A recent statistical study by Fujibe et al. (2009) and several case studies suggest that the urban effect has some type of impact on heavy rainfall in Tokyo, Japan. As a target area of TOMACS, Tokyo is one of the biggest cities in the world, characteristics of the highly urbanized area should be appropriately represented in numerical models.

Numerical simulations for cases of heavy rainfall in Tokyo have been conducted with the JMA nonhydrostatic model (NHM) that incorporates the square prism urban canopy scheme (SPUC) developed by MRI (Aoyagi and Seino 2011). In contrast to the slab land surface scheme used in the operational version of NHM, SPUC takes into account heat and radiation exchanges by the urban canopy elements. Time and spatially varying anthropogenic heating in the metropolitan area is also considered in SPUC. Area fractions of buildings and other land use categories in each model grid are determined from the 100-m-mesh Digital National Land Information Dataset. A6.2. Data assimilation

As a goal of the second theme, the “development of the monitoring and forecasting systems of severe phenomena” is applied to improve the early detection methods of severe phenomena. In this section, the studies in which the numerical models were conducted by MRI are explained.

Although recent rapid progresses in the design and construction of computers have yielded significant improvements in numerical forecasts, quantitative forecasts of severe

20

phenomena, such as local heavy rainfall and tornados, are still substandard. In particular, time, position, and intensity predictions of local heavy rainfall, of which temporal and horizontal scales are small and develop under unstable atmospheric conditions, are difficult to conduct despite recent advances in computing technologies.

To improve the forecasting and monitoring of severe phenomena, accurate 3D distributions of airflow and water vapor are needed. To achieve this, data assimilation methods for high resolution data and high resolution numerical models should be developed. Seko et al. (2007) and Kawabata et al. (2011) conducted data assimilation studies of the Nerima heavy rainfalls that occurred on 21 July 1999, in which high resolution observation data were used. Seko et al. (2007) assimilated radial wind from the Haneda and Narita Doppler radars and GPS-derived Precipitable Water Vapor (GPS-PWV) by using the 3DVAR system of the JMA NHM, and reproduced the low-level convergence of the northerly, easterly and southerly airflows. The importance of convection-scale water vapor was also pointed out. In particular, areas where the updrafts exceeded 0.16 m/s in the rainfall regions were analyzed by using the statistical relationship between updrafts and relative humidity. This modification of water vapor succeeded in the reproduction of the local heavy rainfalls. Kawabata et al. (2007) developed a 4DVAR system with a horizontal grid interval of 2 km for the JMA NHM, and assimilated the radial winds of the Doppler radars and GPS-PWV and assimilated the surface wind and temperature of AMeDAS for every minute, 5 minutes, and 10 minutes, respectively. In their assimilation results, the low-level convergence of horizontal wind near Nerima and the generation of intense convection cells were well reproduced. In these experiments of the Nerima heavy rainfall, the low-level convergence in the non-precipitation echo areas that was obtained by the Doppler radars is one of their success factors.

Kawabata et al. (2011) performed data assimilation experiments of the Shutoken local heavy rainfall that occurred on 4 September 2005 by using the NHM-4DVAR system, in which microphysical processes were introduced. The local heavy rainfall in this case event was caused by a rainfall band that had a width of 15 km and a length of 100 km, which remained at the southern areas of the Kanto area from 20:00 JST on September 4 to 01:00 JST on September 5. They assimilated the Doppler radial winds, reflectivity, GPS-PWV, surface wind and temperature of AMeDAS, and the wind profiler’s wind profiles every minute, 5 minutes and 10 minutes. The assimilation of this high resolution data reproduced the intense rainfall band. Kawabata et al. (2012) also investigated the impacts of radial wind obtained from the Doppler Lidar of the National Institute of Information and Communications Technology, and showed that the inflow modified by assimilation improved the forecasts of thunderstorms (Fig. A6.3).

In the current project, the following were initiated in 2012: observations of rainfall and

21

radial winds obtained from many of the X-band radars and X-NET, observations of radial wind from the Doppler lidars, and observations from a dense network of automated weather stations and a dense GPS array. A detailed structure of the thunderstorms and the distributions around them are expected to be obtained by the observation data and the analyzed results of assimilation systems. To obtain the high resolution distributions, the assimilation methods of the lidars’ radial winds and the slant delays of the GPS dense array have already been developed.

Fig. A6.2. Scheme of the data assimilation experiments (Kawabata et al. 2011).

(a) (b)

Fig. A6.3. Horizontal distribution of 1-h accumulated rainfall amounts (colors) from 16:00 to 17:00 JST and sea level pressure (contours) at 17:00 JST that were obtained by (a) assimilation of the lidar data and (b) a control forecast.

A

B

Fig. A6.1. Data assimilation domain and positions of the observation data. ×, △ and ● indicate the positions of the AMeDAS, GPS-PWV and Doppler radar sites (Kawabata et al. 2011).

22

The results from the data assimilations are fields that are obtained via the likelihood method, which is based on the errors of first guesses and observation data. The results include analyzed errors, even if the high resolution observation data were used. In particular, local heavy rainfall events, which are composed of convection cells developed in weak convergence regions under unstable atmospheric conditions, are sensitive to the initial conditions. Namely, small differences in the initial conditions caused large differences in the forecast results. The influence of errors on the initial conditions can be evaluated with ensemble forecasts. Ensemble forecast systems have been developed by MRI (e.g., Kunii et al. 2011; Saito et al. 2011a), and were used in the WWRP RDP Beijing Olympic project.

To predict local heavy rainfall events, ensemble forecast systems that can reproduce convection scale phenomena should be developed. An ensemble forecast system for mesoscale phenomena has been used in some projects. For instance, Seko et al. (2011) implemented a data assimilation system using the Local Ensemble Transform Kalman Filer (LETKF) based on the JMA NHM (Miyoshi and Aranami 2006), and showed that a convection band, which developed in Kobe City, was reproduced by the assimilation of GPS PWV data in experiments with a smaller grid interval of 2 km. Prior to this experiment, Shoji et al. (2009) performed data assimilation experiments of GPS-PWV by using JMA’s operational data assimilation system (Meso-4DVAR system) to show the impacts of PWV, which was obtained by quasi-real time analysis on rainfall forecasts, and concluded that water vapor increments at the upstream side of rainfall regions affects rainfall forecasts.

To utilize the data obtained by the dense observation network of this project, a nested-LETKF system that reproduces large scale convergences by an Outer LETKF with a grid interval of 15 km and that assimilates high resolution data, such as GPS-PWV and those provided by X-band radars, by an Inner LETKF with a grid interval of 2 km, is under development. The data assimilation experiments using the nested LETKF system will be performed in cooperation with other projects, such as the Strategic Programs for Innovative Research. A7. Social experiments

In 2010, the Japan Science and Technology Agency (JST) launched a new research program called the "Social System Reformation Program for Adaption to Climate Change" under the "Special Coordination Funds for Promoting Science and Technology" of the Ministry of Education, Culture, Sports, Science and Technology (MEXT). TOMACS is one of the research projects selected for the JST research program. TOMACS began in July 2010 with more than 25 organizations and over 100 people participating in the project.

23

The project test-bed for TOMACS is the Tokyo metropolitan area, which is defined in the present study as the 50 km area surrounding the headquarters of the Tokyo Metropolitan Government. The Tokyo metropolitan area is the world’s most populous metropolitan area. The total population is approximately 30 million. About 10 million people use some type of public transportation every morning to commute to their schools and/or offices. The area includes the Tokyo, Kanagawa, and Saitama Prefectures and a part of Chiba Prefecture. The following five mega-cities are within the area: Tokyo (8.5 million), Yokohama City (3.6 million), Kawasaki City (1.3 million), Saitama City (1.2 million), and Chiba City (0.9 million).

TOMACS aims to understand the processes and mechanisms of extreme weather by the utilizing dense meteorological observation networks to develop a monitoring and predicting system of extreme phenomena. In addition, TOMACS aims to implement social experiments on extreme weather resilient cities in collaboration with related governmental institutions, local governments, private companies, and residents.

The social experiments are carried out in the following four fields of disaster prevention: (1) rescue services, (2) risk management, (3) infrastructure, and (4) education. Before implementing a social experiment, surveys were conducted for each experimental field on what information is appropriate and on the effective means of transmitting such information. For rescue services, real-time and nowcast rainfall data derived from a dense radar network are provided to the rescue staff to evaluate how such information can make rescue activity more efficient. For risk management, forecast information is given to the staff of local governments. Warnings based on this information are informed to citizens experimentally. In the infrastructure experiment, warnings of LHIW are provided to a construction site by a siren and email. For educational purposes, high school students are also incorporated into this project. We use observation and forecast data to raise awareness of LHIW to the students.

Within the international framework, we will exchange information on how to provide high-resolution forecast data to the public and how to evaluate its social impacts with related projects such as the CASA Dallas-Fort Worth Experiments.

A8. Liaison with the Dallas-Fort Worth Experiment

The Dallas Fort Worth Water Research Network (DFW-WARN) is a research and innovation network linking academic researchers, local stakeholders, and industry to address water issues as they relate to urban sustainability, in particular, flood hazard warning and mitigation as well as management and design of urban water infrastructure. The DFW-WARN will leverage the remote sensing assets of the CASA Dallas-Fort Worth (DFW) Urban Test bed, a network of eight X-band, dual-polarimetric radars to be deployed across the DFW region linked to existing in-situ observational networks of rain gauges and stream gauges, and existing radars. The network will provide comprehensive,

24

low-altitude mapping of real-time rainfall rates across the metroplex with unprecedented spatiotemporal resolution and increased accuracy (Wang et al. 2010) compared to the state of the art. By linking these new observations with the appropriate cyber infrastructure, existing multi-sensor databases and hydrologic-hydraulic modeling frameworks, we will enable the development of new scientific knowledge. The goal of DFW-WARN is to create new knowledge that translates to user-relevant information for better policy, and organizational and individual decision-making. This effort necessitates a multidisciplinary approach comprised of natural, environmental, computational, social, policy, and cultural perspectives (NSTC 2007).

DFW-WARN will draw together researchers across disciplines and sectors, with outreach to the cities of Tokyo and Mumbai for cross-cultural comparisons of human behavior and exchanges of other scientific and technical research. Primary academic partners include the University of Texas at Arlington (UTA, lead), University of Massachusetts at Amherst (UMass), Colorado State University (CSU), the University of Colorado at Colorado Springs (UCCS), the University of North Texas (UNT), and the University of Wisconsin (UW). Regional stakeholder partners include the North Central Texas Council of Governments (NCTCOG), the emergency management community in DFW, Vision North Texas, the hydrologic ensemble prediction (HEP) Testbed for the Upper Trinity River and the West Gulf River Forecast Center.

A9. RDP preparatory meeting in October 2012

In preparation of the RDP proposal to WWRP, a preparatory meeting was held in Shinagawa Intercity, Tokyo, on 25 October 2012. Approximately 50 participants from six countries attended the meeting. The meeting’s agenda is listed below, which is also accessible on our website (http://www.mpsep.jp/notify/369.html).

0930-0940: M. Maki (NIED), Opening address 0940-0945: K. Sasaki (JMA), Expectations for TOMACS 0945-1000: K.Saito (MRI), WWRP working group and RDP 1000-1200: V. Chandrasekar (CSU, USA), The Dallas-Fort Worth Urban Demonstration Network for

disaster mitigation. A. J.Pereira (SU, Brazil), A hydrometeorological nowcasting system for the metropolitan area

of São Paulo,Brazil. S. Belair, P. Joe, S. Leroyer, S. Pellerin (EC, Canada), Mesoscale urban numerical modeling

for TOMACS. A. Seed (BOM, Australia), SREP radar enhancement project. D.-I.Lee (PKNU,Korea), Orographic precipitation observation in Jeju island, Korea (June 25 -

July 15, 2012). D. Schertzer and I. Tchiguirinskaia (UANPE, France), F.Blachet and B. Tisserand (Veolia

Water), Potential,multiscale contributions to TOMACS and WWRP. S. Sugimoto (CRIEPI) and J.Sun (NCAR,USA), Application of Variational Doppler Radar

Data Assimilation (VDRAS) System to a severe storm occurred in TOMACS campaign area: Initial results.

25

1310-1510: Y. Yamada (MRI), Review of IOP campaign. N. Seino, T. Kawabata, N. Nagumo, T. Aoyagi (MRI), H. Sugawara (NDA), R. Oda (CIT),

urban modeling and radiosonde observation -urban research activities in TOMACS-. Y. Fujiyoshi (Hokkaido Univ.), Monitoring of megacity clouds and wind fields by 3D-scanning

Doppler lidar and other instruments. E. Sato, C. Fujiwara, K. Kusunoki (MRI), Ku-band "fast scan" radar observation of

cumulonimbi. Y. Shoji (MRI), Water Vapor Variation around Tokyo-Bay Observed by Dense Network of

GPS. H. Ymauchi, A. Adachi (MRI), T. Murano (TOSHIBA), MRI C-band solid-state polarimetric

radar. K. Iwanami, S. Suzuki, R. Misumi, T. Maesaka, S. Shimizu, N. Sakurai (NIED), Overview of

the life cycle of cumulonimbus experiment (LCbEx) by NIED. N. Sakurai, K. Iwanami, T. Maesaka, S. Suzuki, S. Shimizu, R. Misumi, D.-S. Kim, M. Maki

(NIED), Cumulonimbus development observed by Ka-band Doppler radar. S. Shimizu (NIED), Data assimilation experiment of thermodynamical parameters retrieved

from high-frepuency updated dual-Doppler analysis data. H. Seko, T. Tsuyuki, K. Saito, M. Kunii (MRI/JMA), T. Kuroda (JAMSTEC), T. Fujita (JMA),

T. Miyoshi (Univ.of Maryland), Data assimilation experiments of convective systems. K. Saito, M. Kunii, S. Saito (MRI), Ensemble experiment of the 26 Aug 2011 mesoscale

convective system. 15:30-17:00: Membership of ISSC, LOC, Milestone to proposal (closed meeting) A10. References Aoyagi, T. and N. Seino, 2011: A square prism urban canopy scheme for the NHM and its

evaluation on summer conditions in the Tokyo metropolitan area, Japan. J. Appl. Meteorol. Climatol., 50, p1476-1496.

Fujibe, F., H. Togawa, and M. Sakata, 2009: Long-term change and spatial anomaly of warm season afternoon precipitation in Tokyo. SOLA, 5, 17-20.

Hirano, K. and M. Maki, 2010: Method of VIL calculation for X-band polarimetric radar and potential of VIL for nowcasting of localized severe rainfall -Case study of the Zoshigaya downpour, 5 August 2008-, SOLA, 6, 89-92.

Ishihara, M., 2013: Radar echo population of air-mass thunderstorms and nowcasting of thunderstorm-induced local heavy rainfalls. Part 1: Statistical characteristics. J. Disaster Research, 8, 56-68..

Kawabata, T., T. Kuroda, H. Seko and K. Saito, 2011: A cloud-resolving 4DVAR assimilation experiment for a local heavy rainfall event in the Tokyo metropolitan area. Mon. Wea. Rev., 139, 1911–1931.

Kawabata, T., H. Seko, K. Saito, T. Kuroda, K. Tamiya, T. Tsuyuki, Y. Honda and Y. Wakazuki, 2007: An assimilation experiment of the Nerima heavy rainfall with a cloud-resolving nonhydrostatic 4-dimensional variational data assimilation system. J. Meteor. Soc. Japan, 85, 255-276.

Kawabata, T., H. Iwai, Y. Shoji, H. Seko, and K. Saito, 2012: Assimilation experiment on a local heavy rainfall event using Doppler Lidar observations, Proceedings of Second

26

International Workshop on Nonhydrostatic Numerical Models. (http://wind.gp. tohoku.ac.jp/nhm2012/program_0919.pdf)

Kunii, M., K. Saito, H. Seko, M. Hara, T. Hara, M. Yamaguchi, J. Gong, M. Charron, J. Du, Y. Wang and D. Chen, 2011: Verifications and intercomparisons of mesoscale ensemble prediction systems in B08RDP. Tellus, 63, 531–549.

Miyoshi, T. and K. Aranami, 2006: Applying a Four- dimensional Local Ensemble Transform Kalman Filter (4D-LETKF) to the JMA Nonhydrostatic Model (NHM), SOLA, 2, 128-131.

National Science and Technology Council (Executive Office of the President of the United States), 2007: A strategy for federal science and technology to support water availability and quality in the U.S., September 2007, 46pp.

Saito, K., M. Hara, M. Kunii, H. Seko, and M. Yamaguchi, 2011: Comparison of initial perturbation methods for the mesoscale ensemble prediction system of the Meteorological Research Institute for the WWRP Beijing 2008 Olympics Research and Development Project (B08RDP). Tellus, 63, 445–467.

Seko, H., Y. Shoji, and F. Fujibe, 2007; Evolution and airflow structure of a Kanto Thunderstorm on 21 July 1999 (the Nerima heavy rainfall event), J. Meteor. Soc. Japan, 85, 455-477.

Seko, H., T. Miyoshi, Y. Shoji and K. Saito, 2011: A data assimilation experiment of PWV using the LETKF system -Intense rainfall event on 28 July 2008-. Tellus, 63, 402–414.

Shoji, Y., M. Kunii, and K. Saito, 2009: Data Assimilation Experiments of GPS PWV in and around Japan on the 28 July 2008 Hokuriku and Kinki Heavy Rainfall. SOLA, 5, 45-48．

Shoji, Y., 2013: Retrieval of Water Vapor Inhomogeneity using the Japanese Nationwide GPS Array and its Potential for Prediction of Convective Precipitation, J. Meteor. Soc. Japan, 91, 43-62.

Wang, Y., and V. Chandrasekar, 2010: Quantitative precipitation estimation in CASA X-band dual-polarization radar network. J. Atmos. Oceanic Technol., 27, 1665-1676.

Webb, F. H., and J. F. Zumberge, 1993: An introduction to the GIPSY/OASIS-II. JPL Publ., D-11088.