Country Report (2004) For the 37th Session of Committee ESCAP/WMO

Shanghai, P. R. 16 – 20 November 2004

People’s Republic of China

I. Overview of Meteorological and Hydrological Conditions during the Year

1. Meteorological Assessment From Jan.1 to Sept. 30 2004, in total 22 tropical cyclones were formed over the Western North Pacific and the (Figure1.1). The total number was well over the average (19.77). 1 TC was formed in April, 2 in May, 5 in June and 8 in August. The number was more than the average 0.73, 1.04, 1.83 and 5.83, respectively. 3 TCs were formed in July and 3 in September. The number was less than the average 4.13 and 3.58, respectively. Six tropical cyclones landed over China, including (0407), Tropical Storm Kompasu (0409), Tropical Storm Nameless (04xx), (0413), (0417) and Tropical Storm Haima (0420). Moreover, there were eight tropical cyclones impacting the off-shore region of China without making landfalls: (0402), Conson (0404), Chanthu (0405), Dianmu (0406), Megi (0415), Chaba (0416), Songda (0418) and Meari (0421).

Figure 1.1 Tracks from Jan. to Aug., 2004

There were four characteristics for the tropical cyclones this year. First of all, they had a longer lifespan. Among all the cyclones, 13 lasted for more than (exclusive) 5 days and they made up 59.1% of the total. The one with the longest lifespan was Chaba (0416) and it lasted for 12 days. Secondly, the source regions of tropical

1 cyclones were relatively concentrated. The active regions of tropical cyclones in the Western North Pacific usually were located from the east of to 150oE. There were 14 tropical cyclones born in this area from January to September this year and they made up 63.6% of all. Furthermore, 5 tropical cyclones generated in the region between 150oE and 180oE and they made up 22.7%. And 3 tropical cyclones formed over the South China Sea, its number is less than the average (3.54). Thirdly, there was an obvious concentrated period of tropical cyclone occurrence. Among the 22 tropical cyclones, 16 were generated between June and August, the number of which amounted to 72.7% of all, which was higher than the average 59.64%. Moreover, 8 tropical cyclones were born in August and they accounted to 36.4% of all, which was higher than the average 29.5%. Finally, the intensities of tropical cyclones were usually strong. Among the 22 tropical cyclones, 15 met the typhoon intensity, and they made up 68.2% of all and this ratio was higher than the average 59.9%. During 2004, the climate backgrounds or the climate conditions, which are all considered as the major factors for the prediction of typhoon activity in 2004, were as follows: SST in the Western Pacific in 2004 was near normal. SSTA in the Western Pacific in 2004 was higher than that in 2003 but lower than that in 2002 (Figure1.2).

Figure1.2. Monthly Mean SST (top) and SSTA (bottom) in July, 2004

„ Tropical convection activity indicated by OLR in the Western Pacific in the 2003/04 (winter) was stronger than that in the 2002/03 (winter), but weaker than that in the 2001/02 (winter). „ The NW Pacific Subtropical High was stronger and more westward than normal. „ The precipitation amount and areas hit by during Jun. - Sep. 2004. There were totally 9 typhoons that impacted China and brought precipitation in land during this period (Figure1.3), with 6 landfall cases (MINDULLE, KOMPASU, an unnamed tropical storm numbered 04XX only by China, RANANIM, AERE and HAIMA) and 3 offshore affecting cases (CHANTHU, MEGI and MEARI). With regard to impacted area, typhoon RANANIM was the most important one, with a precipitation

2 volume of 56km3 which is about 1.5 times of the dam volume of the Three Gorge Dam, and the impacted area was 2,040,000km2 which is more than 1/5 of the area of mainland China. Undoubtedly, RANANIM was the strongest typhoon that hit Province in the past 48 years. In addition, typhoon AERE was the second important case during the period, with a precipitation volume being 21km3 and the impacted area reaching 1,180,000km2, which were much smaller than these of RANANIM.

250 60

impacted impacted area(10000km 台风影响面积 impacted area volume(k precipitation 台风体积降水precipitation volume 50 200

） 40 150 10000km2

（ 30 积 面

响 100 影 m 风 2 ) 20 3 台 )

50 10

0 0 CHANTHU123456789 MINDULLE KOMPASU Nameless RANANIM MEGI AERE HAIMA MEARI 灿都 蒲公英 圆规 (04xx) 云娜 鲇鱼 艾利 海马 米雷 Figure1.3 The precipitation volumes and impacted areas for typhoon cases

that impacted China during Jun. to Sep. 2004

2. Hydrological assessment No basin-scale flood occurred during the flood season of 2004 in China. Some small and medium-scale rivers, such as Lihe river in Huai River Basin, Liujiang River in Pear River Basin, Yuanjiang River and Zishui River in Dongting Lake Catchments, witness serious floods caused by torrential rainfall, especially in Niyanghe River, tributary of border River Brahmaputra, and Dayingjiang River, tributary of border River Yiluowadi, flood crest of which broke the historical records. Among 6 tropical cyclones which landed along the coastline of China in 2004, Rananim was most notable one in term of its exceptionally high intensity, longer duration and large scope it affected. 7 Provinces experienced the influence Rananim posed. The maximum rainfall for 12 hours and 24 hours caused by Rananim in Zhejiang province, reached the record in 100 years. However, the heavy rainfall brought by Rananim relieved the long-time drought to some extent. Thanks to the effective measures to combat flood and drought, especially the timely and accurate hydrological data collection, information dissemination, flood forecasting and so on, China had a relatively smooth and safe flood season basically. 3. Socio-economic assessment

3 The landed typhoons and tropical storms in China brought abundant precipitation, and abated the agricultural drought in the southern Yangzi River, and the reservoir water storage was increased. However, the violent gust, heavy rain and associated astronomical tides also brought about severe losses in the coastal areas during this year, especially in Zhejiang Province. According to the preliminary statistics, 22.58 million people and 8890 km2 farmland were affected by tropical cyclones, and 176 are killed, and left 53 missing and 2126 injured or sick, 85.9 thousands of houses collapsed and 265.1 thousands of houses were destroyed. The direct economic losses were about 23.5 billion RMB Yuan. Comparing the disaster losses with those of the last 10 years, the economic losses caused by the typhoon and storms in China during January to August in 2004 were less severe than the first 8 years but more severe than those of the last two years.

4 II. Meteorology

1. Progress in Member’s and Regional Cooperation and Selected RCPIP Goals and Objectives a. Hardware and/or Software progress

„ The geo-stationary meteorological satellite FY-2C FY-2C is the first operational Geostationary Meteorological Satellite in China, which is located in 105°E, it is also the first one for the second FY-2 group, which is consisted of three satellites. It is planned that this group will be in service from 2004 to 2012. FY-2C was successfully launched in October .19th in 2004, and ten day later, the first visible image was received (Figure2.1), and now it is in orbit test period，it will be in operational service on March in 2005.

Figure2.1 The First Visible Image of FY-2C

Compared with first group of FY-2, the second group has 5 spectral channels on Visible and Infrared Spin Scan Radiometer (VISSR); it means two channels are increased. ¾ The infrared window channel (10.5-12.5µm) is split to two channels. It will improve the accuracy of sea surface temperature (SST) product. ¾ The new channel (3.5-4.0µm) can improve the accuracy of land surface temperature (LST) product, cloud parameters and the watching capability of forest/grass fire.

5 ¾ The visible channel band is changed from 0.5-1.05µm to .55-0.90µm; this will reduce the affection of absorbability to visible data. ¾ The infrared channels radiative resolution are improved, the quantification scale is changed from 8bit to 10 bits. ¾ The space resolution of visible channel is also improved from 1.44km to 1.25km. The characteristics of the FY-2 second group VISSR are shown in Table2.1 and Table2.2. Table2.1 Visible Channel Characteristics Wavelength 0.55～0.90 micrometer FOV 35 microrad Resolution (at Nadir) 1.25 km Dynamic Range of Sensor 0~98% Noise Performance S/N=1.5 ( Albedo =0.5%) S/N=50 (Albedo=95%) Number of Sensor 4 + 4 Quantification Scale 6bit Calibration Solar and Electronic

Table2.2 Infrared Channels Characteristics Band IR1 IR2 IR3 IR4 Wavelength(µm) 10.3～11.3 11.5～12.5 6.3～7.6 3.5～4.0 FOV 140 140 140 140 Resolution(km) 5 5 5 5 Dynamic Range of Sensor 180～330K 180～330K 190～300K 180～340K Tem. Resolution 0.4～0.2K 0.4～0.2k 0.5～0.3Ｋ 0.6～0.5Ｋ Number of Sensor 1 + 1 1 + 1 1 + 1 1 + 1 Quantification Scale 10bit 10bit 10bit 10bit

The products include: ¾ All kind of image, such as full disk image and all sub area image; ¾ Automatic quantitative products, they are TBB, OLR, SST, Cloud Mask, Cloud Wind, Precipitation estimating and index, Albedo, Water vapor, etc. ¾ Interactive products, they are Snow Cover, Sea Ice, Soil Moisture, Sand Dust, Fog, Fire, Water Condition (Flood), Tropical Cycle Location, etc. ¾ Besides the observation, FY-2 Second group satellites have the function of data collection, products transmission and space environment monitoring.

„ Beijing– GTS circuit was upgraded in February, 2004. The upgraded link is connected to the RMDCN of RA VI by using the Frame Relay service provided by EQUANT, and it has a CIR of 48Kbps. TCP sockets are adopted to exchange data on the link.

6 b. Implication to Operational Progress „ To collect domestic aircraft observations, a leased telecommunication line was installed between CMA and CAAC in August, 2002. It is a Frame Relay link, and it has a CIR of 32Kbps. CMA started receiving domestic aircraft observations from November, 2002. At CMA, the domestic aircraft observation data is translated into AMDAR bulletins, and then distributed to the local weather centers in China and the GTS centers connected with RTH Beijing. The routine distribution of domestic AMDAR bulletins started from October 1, 2004. c. Interaction with users, other Members, and/or other components

Nil. d. Training Progress

6 training courses on Doppler weather radar were held in CMA Training Center in 2004, with total number of the participants being about 300. For the first time as the Beijing component of WMO Regional Meteorological Training Center Nanjing, it enrolled two participants from the Hydro-Meteorological Service. Some lectures in the training focused on how to use radar to locate the center of typhoon, estimating the intensity (wind speed), estimate the typhoon precipitation and warning techniques of severe convective weather associated with typhoon spiral rain band. CAL CD-ROM on Doppler radar developed by CMA Training Center was given to the Vietnamese participants as a reference. e. Research Progress

Nil. f. Other Cooperative/RCPIP Progress WMO/ESCAP Typhoon Committee Roving Seminar will be held at the National Meteorological Centre, China Meteorological Administration in Beijing from 22 to 24 November 2004. The seminar will focus on Operational application of multi-model ensemble typhoon forecasts. Dr. Nobutaka Mannoji from JMA and Dr. Russell L. Elsberry from CAS have been invited as speakers at the seminar.

2. Progress in Member’s Important, High-Priority Goals and Objectives a. Hardware and/or Software progress „ 582 Automatic Weather Stations (AWSes) became operational from Jan. 1, 2004, which were established under the Automatic Atmosphere Monitoring System project. Thus, in total 1606 AWSes have been operational up to now.

„ From Jan. 1, 2004, the new Surface Meteorological Observing Practice was implemented at all surface observing stations operated by the China Meteorological

7 Administration (CMA), which is applicable not only to the automatic observation but also to the manual observation. In compliance with the new practice, the new Operational Software of Surface Meteorological Observation (OSSMO) has been developed as well. The new OSSMO version has been tested and will become operational next year. „ In 2004, the capability building of upper air sounding network in China was continuously enhanced. 10 sets of L-band secondary wind-finding radar & digital sonde system have been put into operation as from August 1, as a result, the quality of sounding data at sites has been obviously improved. „ According to the upgrated plan, within the China New Generation Doppler Radar Network will consist of 158 new-generation CINRAD doppler radars. So far 64 CINRAD radars have been already installed in China and successfully used in observing hail&rainstorm or monitoring typhoons. It has made substantial contribution to the disastrous weather preparedness and mitigation in many provinces. In Southeast China, the radars in Zhoushan, Changle and the monitored typhoons successfully and led to accurate forecasts. They have made excellent contribution to monitoring of disastrous weather events in Zhejiang and provinces. Till the end of this year, altogether about 93 CINRAD radars will deploy in mainland China. We plan to install 29 or so additional CINRAD radars next year. „ Up to September 2004, National Satellite Metrological Center (NSMC) has monitored and analyzed 22 TCs with the help of meteorological satellites. The tasks include locating TC center, intensity estimating, track monitoring, structure analyzing and so on. Also QUIKSCAT sea surface wind vector and AMSU microwave data are used to analyze the regional distribution of the gale and the heaviest rainfall. NSMC releases the real-time results of the analysis of TCs, such as the location of TC, its intensity, its track and synthetical result with the multi-data on the website www.dear.cma.gov.cn. b. Implication to Operational Progress

„ The Global Model for Typhoon Track Prediction (GMTTP), coupled with global spectral model T213L31, was put into quasi-operation in the 2004 typhoon season. ¾ Vortex specification The initialization of GMTTP model still adopts bogus vortex scheme used in The Regional Model for Typhoon Track Prediction (RMTTP), but some modifications have been made, such as, the calculation of troposphere based on the global model atmosphere conditions instead of an original fixed value and the way to form background fields of typhoon vortex.

8 z To remove ill-defined and weak vortex from the large-scale background fields in order to maintain smooth environmental field.

z To construct a symmetric typhoon vortex based on a few parameters manually analyzed by forecasters: the TC central position (latitude, longitude), the central pressure and the maximum wind speed radius.

z To merge the synthetic typhoon vortex into the smooth environmental field by a blending method with a linear weighting function. ¾ GMTTP Forecast Model The GMTTP model is coupled with global spectral model T213L31, so it includes all the T213L31 characteristics. 0 0 z Horizontal resolution: 0.5625 X 0.5625

z Vertical resolution: 31 levels

z Forecast domain: East Asia, Northwest Pacific Ocean ¾ Guidance products The GMTTP system runs twice a day and provides every 6 hour TC center location and intensity forecast in the 72 hour time period at 00 UTC and the 96 hour time period for 12UTC respectively. ¾ Verification A parallel test on the 2003 typhoon cases between was made the new GMTTP and the old RMTTP was carried out in order to verify the prediction capability of the GMTTP model. The statistical analysis of forecast position errors indicates the GMTTP model performs better than the old one (Figure 2.2). It shows a reduction in the average track error of about 100km in the 48 hour time period.

500 450 RMTTP 400 GMTTP 350 300 250 200 150 100 mean track error (km) 50 0 0 1224364860728496 Forecast period (hour)

Figure2.2 Comparison of mean track error between GMTTP and RMTTP in 2003

9 In 2004, the GMTTP was put into quasi-operation instead of RMTTP. During the typhoon season, the new system operates well and can offer real-time products to forecasters. It shows the typhoon Mindule (0407) forecast tracks as Figure 2.3.

Figure2.3 Tracks of Typhoon Mindule (0407) predicted by GMTTP

In 2004, the verification of GMTTP mean forecast track error is shown as Figure2.4 and table2.3.

500 450 400 350 300 250 200 150

mean track error(km) 100 50 0 12h 24h 36h 48h 60h 72h 84h 96h Forecast period

Figure2.4 The Mean Track Errors for GMTTP (unit: km)

10

Table2.3 Mean Track Errors for GMTTP (unit: km) Forecast 12hour 24hour 36hour 48hour 60hour 72houtr 84hour 96hour Period Mean track 90.0 148.9 206.5 264.6 314.6 368.7 404.5 443.6 errors (214) (198) (182) (166) (151) (137) (62) (55) Note: the figures in the brackets are the numbers of forecasts. Compared with the RMTTP model, the GMTTP model makes progress in forecasting track error during the 2004 typhoon season. The detailed results are shown in Figure2.5.

500 450 RMTTP 400 GMTTP 350 300 error (km) 250 200 150 100 mean track 50 0 0 1224364860728496 Forecast period (hour) Figure2.5 Comparison of mean track error between GMTTP and RMTTP in 2004

Figures2.5 demonstrates a decrease in the mean track errors for the 48-hour forecast. The improvements illustrate that, to some extent, the new GMTTP model performs better than the old RMTTP model at every time period for the 2004 typhoon season. „ The reductions of the mean errors of the subjective operational forecast for tropical cyclones track in NMC/CMA during the past decades

Since 1991, especially over the recent 4-5 years, the mean errors of subjective operational prediction for tropical cyclones track in NMC/CMA have been reduced gradually. During the past three years, the typhoon-track forecasting errors within 24 hours were below 150km while the 48h forecast errors were below 250km (Figure2.6, 2.7). The NMC began to issue the 72h typhoon-track and intensity forecasts in 2001, and its mean forecasting errors within 72 hours were 419km, 364km and 365km respectively from 2001 to 2003, making noticeable achievements.

11 250

200

150

100

50

0 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Figure2.6 The 24h Track Mean Errors of subjective operational prediction for tropical cyclones in NMC/CMA since 1991 (km)

600

500

400

300

200

100

0 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Figure2.7 The 48h Track Mean Errors of subjective operational prediction for tropical cyclones in NMC/CMA since 1991 (km)

„ Using multi-satellite, multi-spectrum channels to synthetically analyze TCs in NSMC.

During these years, NSMC uses multi-satellite and multi-spectrum channels to synthetically analyze TCs. The satellites in use include: ¾ Stationary satellite FY-2B、GOES-9 ¾ Polar-orbiting satellite FY-1D、NOAA-12、NOAA-16、EOS ¾ Sea surface wind vector monitoring satellite QUIKSCAT Besides the regular monitoring and analysis of TCs, the group of TCs in NSMC has mainly finished 4 tasks. They are as follows:

12 ¾ Making the satellite annuals of the TCs in 2003. Based on the satellite analysis of the TCs in 2003, a CD named the annuals of the satellites observed TCs in 2003 has been prepared. The CD contains the location, track, animated cloud images of all TCs in 2003. Also satellite data such as QUIKSCAT wind vector, microwave, strengthened cloud images and related retrieval satellite data are used to analyze the structure, impact factor and development of the TCs. ¾ Analyzing the structural characteristics of TCs with the polar-orbiting satellite Polar-orbiting satellite can provide high resolution images about TCs. FY-1D、NOAA series satellite nadir resolution is 1.1km. The resolution of the MODIS sounder on EOS satellite is up to 250m. And the polar-orbiting satellite has more spectrum channels than the geostationary satellite. Based on the data from them, we can look into the circulation characteristics of the TC and the minor structure of the center of the TC. And we can also get more accurate information to estimate the intensity of the TC. Compared with the data from the geostationary satellite, we can get more precise location of the TC. ¾ Estimate the distribution of the area with the heaviest rainfall of TC using the AMSU microwave data AMSU is advanced microwave sounder unit onboard NOAA series satellite. It has 20 channels and can monitor the location, moving, thermodynamic anomaly, velocity and rainfall effectiveness of TC simultaneously. NSMC mainly uses it to analyze the distribution of the area with the heaviest rainfall of the TCs in 2004. The windows of the AMSU-B are very sensitive to raindrops in the cloud. The brightness temperature drops as the number of the raindrops rises. Low brightness temperature of the channel has a good relation with the area of the heaviest rainfall. In practice, we use low brightness temperature region overlap the image of geostationary satellite to know the distribution of the area with the heaviest rainfall. The results play an important role in monitoring the landing typhoon RANANIM (0413) and AERE (0417). ¾ Continue using the QUIKSCAT data to analyze the location and intensity of TCs Based on the tasks finished in 2003, we continue using the QUIKSCAT data to analyze the location and the maximum velocity of the center of TC to supplement the results obtaining from the geostationary satellite. c. Interaction with users, other Members, and/or other components As in stead of old Regional Model for Typhoon Track Prediction (RMTTP), GMTTP can produce tropical cyclone track and intensity prediction and transmitted

13 through GTS twice a day (00, 12UTC) for up to three target TCs at the same time over the western North Pacific. During this typhoon season, the GMTTP has been used into operational tropical cyclones forecasting and gained relatively better results. At the same time, Concensus tropical cyclone track prediction issued by Typhoong Institute also has been submitted to related forecasting of CMA. Figure2.8 and Figure2.9 give the examples of Typhoon Rananim (0413).

Figure2.8 Tracks of Typhoon Rananim (0413) predicted by GMTTP

Figure2.9 Tracks of Typhoon Rananim (0413) predicted by Shanghai Typhoon Institute

At 12UTC of 12 August 2004, Typhoon Rananim (0413) landed over Wenling, Zhejiang province with a maximum wind speed near the typhoon center reaching force 12 (45m/s) when it landed. The NMC made accurate forecasts on the landing time, location and the associated adverse weather events two days in advance (Figure2.10), the timely forecasts and warnings were immediately delivered to relevant governmental agencies, media and the general public. On 12 August, based on the accurate forecasts, Dr. Qin Dahe, Administrator of CMA, personally presented

14 the urgent warnings of the landing typhoon Rananim by in CCTV-1 Weather Forecast program, which is the most popular TV program in China. It produced very positive feed back from media like CCTV etc. and from social communities across the country as well. According to forecasts and warning services before the typhoon landfall, the governments at different levels had enough time to evacuate 467,900 people, and bring 9,900 ships at sea back to harbors or shelters. The casualties and economic losses were reduced significant. The death-tolls caused by Typhoon Rananim were much less than those caused by earlier typhoons (numbered as 7504, 8506, 9417 and 9711 respectively), which choose to make their landfall in Zhejiang province. The typhoon forecast and warning service proved to be satisfactory.

(0413)

Figure2.10 Track and subjective forecast of Typhoon Rananim (0413) d. Training Progress A training course on weather satellite imagery interpretation in weather forecasting was held in CMA Training Center in 2004. Typhoon locating, intensity estimation and the interaction between typhoon and mid-latitude weather systems were discussed in the course. There were 60 participants in the training course. e. Research Progress The research work carried out in the past year focused on the aspects including tropical cyclone structure, intensity change, sustaining and decaying mechanism over land, extratropical transition (ET), track forecasting techniques, application of GRAPES model and climate statistics etc. „ Research activities at the Chinese Academy of Meteorological Sciences

15 The research work carried out by (CAMS) has been focused on the following aspects:

¾ Application of GRAPES Model on Typhoon Simulation Global/Regional Assimilation and Prediction System (GRAPES), based on three dimension and four dimension variational assimilation technique, is being developed by the Chinese Meteorological Administration. GRAPES adopts semi-implicit scheme, semi-Lagrange method and fully compressible non-static equilibrium equations with optimized and free-assembled parameterized schemes. A serial numerical experiment was performed to investigate the effects of GRAPES model on typhoon simulation. Results show that the higher the model’s resolution is, the clearer the typhoon spiral structure will be. The spiral cloud system and worm core structure of typhoon is demonstrated clearly under 10km resolution and the variation of spiral structure and precipitation cloud zones may be distinguished under 5km resolution, but the model will not be able to simulate typhoon with 280km resolution (Figure2.11).

Figure2.11 A comparison of vapor distributions on 600hPa integrated by GRAPES for 24h

Statistics on typhoon prediction during 2003 year indicated that the mean error of typhoon central location was 156KM for 24h forecasting and 248KM for 48h. Besides, GRAPES model is utilized employed to assimilate radar data and satellite TRMM data in typhoon simulation. It was found that the simulation effects on typhoon’s track and precipitation were improved obviously.

16 ¾ Tropical cyclone structure The SNR (signal-to-noise ratio) data of the wind profiler associated with a severe tropical storm Vongfong making landfall in province demonstrated that the turbulence developed up to more than 5km several hours before landfall whereas receded back to 1-1.5 km several hours after landfall. The variation of the boundary layer thickness was coincided with the turbulent layer variation. With QuikSCAT data, the features of surface wind distribution of Vongfong landfalling process were analyzed. It was found that the near-surface wind field was highly asymmetric which changes with the large-scale environmental and typhoon self-rotating circulation. When Vongfong approached coast, the associated gale appeared in front portion in the advancing storm. Besides, Vongfong’s precipitation distribution and convection were also asymmetric based on high-resolution satellite TRMM data. Before the landfall, severe convective cloud bands evolved into spiral cloud bands and precipitation increased while it was rapidly decreased during the landfall. Using power-spectrum analysis and TBB time-serial data, the wave-train characteristics of typhoon’s spiral band were analyzed. Results showed that the TBB departure distribution resembled to the vortex Rossby-wave train and its path was also similar to Rossby wave propagation path. ¾ Tropical cyclone intensity change Under two types of initial tropical cyclone structures that were characterized by high and low vorticity zones, four sets of numerical experiments were made with barotropic model equation to investigate the interaction of a typhoon and an adjacent mesoscale vortex (MSV) and its impact on the tropical cyclone intensity change. The results suggested that the interaction of the tropical cyclone vortex of different structure with the MSV led to different results. Under the condition of low vorticity zonal structure, the interaction led to a pressure decrease of 3.8 hPa at the tropical cyclone center while under the condition of high vorticity zonal structure, the interaction resulted in a pressure decrease of 14.1 hPa. The latter is 3.7 times higher than the former. Therefore, the tropical cyclone structure also played an important role in the intensity change of tropical cyclone itself. Eight numerical experiments were conducted in 48 hours of integration by using a barotropic primitive equation model with a topography component so as to investigate the effect of topography on the merging of vortices. Results indicated that the introduction of topography might change the track of vortices, and it caused the low vortices and vorticity lumps to be detained on the

17 southeast side of the topography, thus creating a favorable condition for the merging of the low vortex and vorticity lumps. It also showed that the effect of topography might cause merges of double vortices in a horizontally shearing basic flow and it could strengthen the low vortex remarkably. ¾ Sustaining and decaying mechanism of a tropical cyclone over land Dynamical and diagnostic analyses were made with composite data of two sets of landfall typhoons (quick dissipation and longer sustention). The results exhibited that the lower layer moisture channel connected with a landing tropical cyclone moving toward north and gains the baroclinic energy from the mid-latitude trough as well as the upper level envolope of the cyclone linking with a mid-latitude jet to enhance the upper level out flow towards northeast which would favorable to sustaining of the tropical cyclone over land. A case study shows that typhoon Bilis (2000) sustained longer period over land with plenty of moisture transported into it. The numerical simulation exhibited that Bilis would rapidly decrease if there was no moisture transportation. The sustention and intensification of typhoon Nina (1975) over land were simulated with MM5V3 and its bogus scheme. The results suggested that the fluxes of latent and sensible heat were favorable to Nina sustention and intensification whereas the flux of kinetic energy would dissipate and filled up the Nina’s depression over land. ¾ Tropical cyclone track prediction The experiences and observations indicated that transition of synoptic patterns surrounding a tropical cyclone was one of the key factors to determine where a tropical cyclone would move in future. The study presented a technique that may be able to identify transition of synoptic patterns so that a tropical cyclone environment could be characterized into “Tropical Cyclone Resistant” or “Tropical Cyclone Favoured”. The study found that the area would depend on the distribution of the atmospheric stratification. The unstable channel would be favorable to tropical cyclone passing through. On the contrary, the stable stratification area would be resistant to the tropical cyclone move towards the area. ¾ Climate statistics The climatic features of the tropical cyclones (TC) influencing China were analyzed by using the TC data from 1956 to 2000. It was found that there existed two closely correlated areas between the annual TC numbers influencing China and the SST anomalies in the Pacific ocean, i.e. the Western Pacific pool region（120~150E, 10~20N）with positive correlation and the central and eastern equatorial Pacific (180~90E, 10S~5N) with negative correlation.

18 Besides, in El Nino years there were fewer but more intense TCs effecting China, but vice verser in La Nina years. With the data from Tropical Cyclone(TC) Yearbooks data (1970 to 2001), statistical analyses were made to study the climatic features of TCs which made landfall on Chinese coastal area, including its landing process, sustaining, decay, transformation, intensification and dissipation etc. Results indicated that the durations of TCs over land were quite different in terms of variable landfall spots. The most obvious decreasing of TCs occurred in the first 12hr after landfall and the stronger a TC was，the more quickly it decayed. Statistical study showed that tropical cyclone annual numbers in Western North Pacific was related to the South Asia High (SAH). The numbers would be more than the mean when the center of SAH went northwestward further and its intensity was lower.

„ Research activities at Shanghai Typhoon Institute The research work carried out by the Shanghai Typhoon Institute has been focused on the following aspects: ¾ GRAPES-Tropical Cyclone Model A “moving-vortex” technique was introduced to the new-generation numerical prediction model GRAPES to set up a GRAPES-Tropical Cyclone Model (GRAPES-TCM). In the experiments for 2002, the average 24h and 48h errors of TC track forecasting were 130km and 252km respectively (Figure2.12). A Bogus scheme based on NCAR-AFWA BOGUS was also introduced, which improved the intensity simulation of TCs effectively.

Average Track Errors of GRAPES_TCM for 2002

300

200

100

track error(km) 0 6 12182430364048 time(h)

Figure2.12. Average track errors of GRAPES_TCM for 2002

¾ Sea-Land-Atmosphere Coupled Model for tropical cyclone study and forecasting An air-sea coupled model for tropical cyclone study was proposed with MM5V3 the atmospheric part and multi-level POM the oceanic part. Idealized

19 experiments demonstrated the model’s capability in simulating tropical cyclone and reflecting the interaction between ocean and tropical cyclone. Using the air-sea coupled model, a series of experiments were carried out to study the effect of changes in the ocean on tropical cyclone intensity and track (Figure2.13).

Figure2.13 The change of SST for the effect of tropical cyclone 8911

The effect of land surface processes was studied by using the Noah Land Surface Model of MM5 simulating the landfall process of an ideal typhoon. Results indicated that the location of precipitation belt was changed remarkably in the experiments with and without the land surface processes (Figure2.14, 2.15, 2.16, 2.17).

Figure2.14 The numerical test of effect of typhoon on sea temperature in different depth

20 Figure2.15 The numerical test of effect of typhoon

on sea stream in different depth

Figure2.16 The difference of 48h accumulated precipitation between

adding meadow and land surface model and only adding meadow

21 Figure2.17 The difference of 48h accumulated precipitation between adding dry surface and land surface model and only adding dry surface ¾ Satellite data assimilation in tropical cyclone simulating and forecasting The three-dimensional varational data assimilation scheme (3DVAR) in the mesoscale model version 5 (MM5) was used to study the effect of assimilating the sea-wind data from QuikScat on the prediction of typhoon track and intensity. The case of Dujuan was first tested and the results showed encouraging improvements. Twelve other cases in 2003 were then further evaluated. With the assimilation, the 24-h (48-h) track prediction of 11 (10) out of the 12 typhoons was improved. The 24-h (48-h) prediction of typhoon intensity was also improved in 10 (9) of the 12 cases. These experiments demonstrated that assimilation of the QuikScat sea-wind data should increase the prediction accuracy of the track and intensity for typhoon through modification of the initial fields associated with the typhoon (figure2.18, 2.19, 2.20).

Novar – Values of Sea Level Pressure Var –Values of Sea Level Pressure (Time:0h)

Figure2.18 The effection of QuikSCAT data assimilation on the initial sea surface air pressure of

22 Figure2.19 Horizontal winds speed of 500hPa assimilation (the left) and non-assimilation(the right) at 083112UTC

Figure2.20 The potential height of 500hPa assimilation (the left) and non-assimilation(the right) at 083112UTC

¾ Tropical cyclone bogus technique Two bogus schemes, NCAR-AFWA and bogus data assimilation (BDA), were compared by carrying out 41 experiments for 9 tropical cyclones in 2002. Results showed that both schemes work to improve the track forecasting with BDA to a greater extent. ¾ Concensus tropical cyclone track prediction techniques A consensus forecast technique based on canonical correlation analyses was proposed for tropical cyclone track forecasting. The three sub-methods selected were official forecasts from China Meteorological Administration (CMA), Meteorological Agency (JMA) and the Joint Typhoon Warning Center (JTWC). Both the developmental and dependent results showed that the concensus output was generally better than any of the three sub-methods. Especially, the 24-h forecast error of moving speed was decreased. ¾ Tropical cyclone intensity Temperature retrievals from NOAA-15 AMSU-A from July to September in 2001 and 2002 were used to develop an estimative algorithm for the intensity of

23 TC in the western North Pacific. The variance (R2) given by the final estimative algorithm was comparable to the published results for the Atlantic and Eastern Pacific TCs. Favorable large-scale conditions were searched for abrupt intensification of TC in Sea before landfall through both case and composite studies. Evidence indicated that, different from the situation in the South China Sea, few common features of the large scale atmospheric environment could be obtained. However, the underlying SST was generally (~ 2° C) higher than those for filling cases, which could be assumed to be a good predictor along East China coast. Such a result indicated that large-scale factors might have played a quite different role in the pre-landfall intensification of TCs in different regions. ¾ Tropical cyclone climatology Differentiating rules were set up to re-build typhoon data from historical disaster records. According to these rules, typhoon data were retrieved for Shanghai City and East China, which could serve as valuable sources for climate study (Figure2.21).

Figure2.21 The review of typhoons causing disaster in Shanghai in the past 200years f. Other Cooperative/RCPIP Progress Nil.

3. Opportunities for Further Enhancement of Regional Cooperation

A number of international conferences regarding tropical cyclone will be held in China in from 2004 to 2005. The Typhoon Committee session will be held in Shanghai from 16 to 20 November 2004. The international workshop on tropical cyclone landfall research is going to be held in Macao, China from 22 to 25 Mar. 2005, and WMO workshop on the interaction between tropical and mid-latitude weather systems will be held in China from 23to 26 Nov. 2005. Those international meetings will be organized by WMO/CAS Working Group on Tropical Meteorological Research. That will benefit regional cooperation in this field.

24 III. Hydrology

1. Progress in Member’s Regional Cooperation and Selected RCPIP Goals and Objectives At the Workshop on Integration of Risk Analysis and Management of Water-related Disasters into Development Process in the Typhoon Committee Area held in , the Philippines in 2002, the representative of China restated to take a major role in implementing 2 of 11 activities (shown in table3.1) identified by the TC Hydrologists in Thailand in the year 2001, based on the RCPIP promises. One was Extension of Flood Forecasting Systems to Selected River Basin. The other was the Project on the Evaluation and Improvement of Hydrological Instruments and Telecommunication Equipment. Table3.1 The activities identified by the TC Hydrologists No. Activity Lead member

Assessment of national requirements and capabilities on hydrological and Review mission 1 Disaster Prevention and Preparedness (DPP) components (concluded) Pilot project for data sharing between TC Members to enhance flood 2 Every member forecasting accuracy. 3 Development of guidelines for the dam operation in relation flood forecasting. 4 On-the-Job Training on Flood Forecasting between TC members Malaysia

5 Extension of flood forecasting systems to selected river basins China

6 Pilot project on the preparation of Inundation and Water-related Hazard Maps Japan Project on the evaluation and improvement of operational flood forecasting 7 Korea system focusing on model performance Pilot project on the establishment of community-based flood forecasting 8 Philippines system Pilot project on the establishment on flash-flood warning system (including 9 Japan debris flow and landslides) 10 Improvement of Hydrological products in response to user needs Philippines Project on the evaluation and improvement of hydrological instruments 11 China and telecommunication equipment China has made some progress during year 2004 under the framework of the Regional Cooperation Programme Implementation Plan (RCPIP) for Hydrology, especially on the projects led by China. a. The Status of two Projects led by China “Extension of Flood Forecasting Systems to Selected River Basin” and “Evaluation and Improvement of Hydrological Instruments and Telecommunication Equipment”, have witnessed some noticeable progress by delivering detailed questionnaires to participating Members, collecting Members’ responses and analyzing the situation of projects application and capacity of offering assistance. At the request of the meeting, China prepared a questionnaire to undertake a

25 survey of the feasibility about the two projects led by China respectively. The questionnaires were distributed to every TC member’s right after 2002 Manila Workshop. In order to get adequate information, the questionnaires were sent out again and called for responses from all TC members in 2003 Beijing Workshop.

„ Extension of Flood Forecasting Systems to selected river basin ¾ Questionnaire and responses The questionnaire for the project on the Extension of Flood forecasting System to Selected River basin, included following overall objectives: z Details about Flood Forecasting Organization in TC members; z Brief introduction of flood disaster happened in TC members; z Information about the Flood Forecasting System existed in TC members; z Interests of participating in the above project. Five TC members China, Japan, Malaysia, Philippines and Thailand replied the questionnaire. ¾ Status of application of flood forecasting system According to the questionnaires, the basic situation of establishing and application of flood forecasting system was realized. All above 5 members have established their own flood forecasting system and they have abundant experiments of model application, especially Japan, Thailand and China. The main methodology actually used by above 5 members in real-time flood forecasting includes hydrological method and experiential method. Besides those two methods, Thailand applied hydraulics method. China has more than 10 models actually used in real-time flood forecasting because it faces all types of flood in its huge territory. The models include Correlation Methods, Xin-AnJiang Model, Excessive Infiltration Model, Antecedent Precipitation Index (API), Maskingum Channel Routing Method, Lag/K Routing Method, Unit Hydrograph Method, Sacramento Model, Tank Model, Storage Function Model and etc. China edited Practical Flood Forecasting Method for each river basin respectively. China Flood Forecasting System has been put into use since 2002 and it is in the process of further improvement. Japan used Storage Function Model, Correlation Methods etc. Japan has a well developed data collection system. A river information system has been developed to communicate river information promptly and accurately that can be used to make sound judgments based on composite conditions in widespread areas. Japan has established an integrated real-time flood forecasting system based on data collection system. Malaysia used Tank model in real-time flood forecasting as well as black box models, such as Stage Regression, Multiple Correlation and Linear transfer function etc. Philippines used simple correlation method, modified Sacramento

26 model etc. Thailand used Mike 11 FF and Mike Flood Watch model in real-time flood forecasting. Fuzzy Neural Network Model is also used in Thailand. ¾ Capacity of offering assistance China, Japan and Philippines express that they have capacity of offering assistance to TC members in the field of flood forecasting. China can provide the guidance on flood forecasting system establishment and forecasting models. China will give a demonstration of simple flood forecasting system. Japan can provide TC members with the information of flood forecasting and warning method and experience in Japan. Philippines like to share their experiences on the applications of models used, share data and etc. ¾ Interests in the project All above countries expressed their interests to take part in the project on Extension of Flood Forecasting System to Selected River Basins. Malaysia chose Timah Tasuh Dam with area of 190 km2 in Perlim river basin as pilot area. Timah Tasuh Dam is a very important dam which has multiple functions of flood control, irrigation and water supply. As the dam catchment frequently experiences heavy rainfall, the need for timely and reliable forecasting is indispensable to reduce flood impact in the downstream areas. Malaysia hopes to develop of a system (such as National Flood Forecasting System in China) which could perform data processing, auto calibration of a forecasting model. The forecasting model could forecast the dam inflows (and reservoir level) and downstream river levels. Hydrologic and hydraulic modules may be needed. Graphical user interface is preferred. Philippines chose Arayat, Pampanga in Pampanga river basin with area of 9000 km2 as pilot area. The area is situated somewhere in the middle of the basin and is a good indicator of flooding in the lower Sections of the basin. Thailand chose Loei Province in Loei river basin with area of 3101 km2 as pilot area. The urban area of Loei province was seriously flooded for several days in 2002. The drainage to the Mekong River was quite slow due to the narrow channel capacity. The flood forecasting and warning system will be very helpful for the residents. ¾ Conclusion From questionnaires, it indicated that development and application of flood forecasting system was uneven. TC members are in urgent need of and are eager for further promoting flood forecasting in aspects of development of system and application of models. TC hydrological working group should push ahead the cooperation and exchange technology and information in the field of real-time operational flood forecasting.

„ Evaluation and Improvement of Hydrological Instruments and Telecommunication Equipment

27 ¾ Questionnaire and its responses The questionnaire for the project on the evaluation and improvement of hydrological instruments and telecommunication equipment includes following overall objectives：

z Situation of the hydrological instruments and telecommunication equipment; z Needs for hydrological instruments and telecommunication equipment; z Interest in the above project; z Capability of offering help to the members on this matter and so on. The questionnaire was replied by China, Japan and Malaysia. ¾ Situation of the hydrological instruments and telecommunication equipment z Data collection methods The hydrological data collection methods adopted generally by TC members includes stationary gauging, tour gauging, telemetry and satellite. Japan applied Radar for data collection. China used GPS, ADPC and PC flow measuring system at some important river stations. z Data transmission methods Japan used digital microwave telecommunication network and Wide Area Network (WAN) for data transmission. Malaysia used public telecommunication network, short-wave and ultra-short wave network and satellite telecommunication network. And China established a mixed data transmission system including all above means in order to meet the need of hydrological observation stations located in different terrain. z Instruments for rainfall and evaporation Instruments for rainfall and evaporation used in TC members include Telemetric Rain Gage, Stainless Steel Tipping Bucket Rain Gauge, Telemetric Evaporation Gauge, Flow Face Evaporation Gauge, Glass Fiber Reinforced Plastic Thermometer Screen and Others. z Instruments for flow current and discharge The instruments for flow current and discharge used in TC members include Propeller Type Current Meter, Velocity Counter, Muddy Water (Sewage) Flowmeter, Radar Current Meter, Portable Current Meter (electromagnetic), ISC Ultra-sonic Open Channel Discharge Meter, Automatic Multi-Control Platform for Cable Hydrometry, Bag Suspended Sediment Sampler, Float measurement, Radio Current meter and others. z Sensors and gauges for water level and water depth acquisition instrumentation z Sensors and Gauges for Water Level and Water Depth Acquisition Instrumentation used in TC members include Absolute Mechanical Coding Water Level Gauging, Bubbler Water Level Gauge, Portable

28 Ultra-sonic Water Depth Gauge, Rainfall / Water Level Data Taker, Telemetric Pressure Water Level gauge, Float type, Reed Switch type and others. z Automatic System of Hydrometry and Hydro-regime data Acquisition System is composed popularly of three parts: Master Station, Repeater and Telemetric station. Telemetric parameters include Rainfall, Water level, Lock lift and Evaporation. Data reporting includes 3 modes: self-reporting, interrogation and compatible. Communication channel used in TC members maybe UHF, VHF, Satellite, PSTN/wireless or Mixed. ¾ Capacity of offering assistance China and Japan expressed their desire to offer assistance to TC members with the information of hydrological observation method and advice on improvement of hydrological instruments and telecommunication equipment. ¾ Interests in the project Both China and Malaysia expressed that they had interests in Automatic System of Hydrometry and Hydro-regime Data Acquisition to promote water information service system. Japan expressed that they had interests in Instruments for Flow Current and Discharge, Sensors and Gauges for Water Level and Water Depth Data Acquisition Instrumentation and Instruments for Rainfall and Evaporation. ¾ Conclusion The review on the Evaluation and Improvement of Hydrological Instruments and Telecommunication Equipment is a preliminary report because of only three countries’ information was received. However, It is obvious that Japan has most advanced hydrological instruments and telecommunication equipment, China and Malaysia needs to further improved their hydrological instruments and telecommunication equipment, but they have their own experience in this aspect. b. Four Projects Participated by China „ Flood Hazard Mapping Project Satisfactory achievements were made in flood hazard mapping project in China. Chinese side strived to fulfill tasks scheduled in the implementation plan on this project in the pilot area, Zhejiang Province, in southeast coastline of China, where often suffer severe floods induced by heavy rainfall or typhoon, and made substantial excellent progresses in aspect of drawing-up of Historical Flood Map, building up Flood hazard Map Management System and setting Loss Evaluation Model to analyze flood loss. ¾ Drawing-up of Historical Flood Map Flood Hazard Maps for 98 cities,70 towns,8 main basins and 4 main plains， had been produced based on basic data of pilot area including topographic map, historical floods in flood inundated years, demographic and socio- economic

29 details. Figure3.1 shows the Flood Hazard Map of Zhejiang Province.

Figure3.1 Flood Hazard Map of Zhejiang Province

¾ Creating a Flood hazard map management system Figure3.2 shows the interface of Flood Hazard Map Management System.

Figure3.2 Interface of Flood Hazard Map Management

¾ Establishing Loss Evaluation Model Figure3.3 shows the interface of Flood Loss Evaluation Model.

30 Figure3.3 Interface of Flood Loss Evaluation Model

¾ Working Plan China also made efforts to develop the future plan in collecting information of relocation area for evacuated dwellers, determining relocation scheme according to the state flood control plan and and incorporating flood prediction in flood hazard map system as next step.

„ Sediment Disaster Forecasting and Warning System China has conducted careful and in-depth preparatory work for the project on sediment disaster forecasting and warning system, which goes smoothly. Substantial achievement has been made in preventing sediment disasters and more attention has been paid to the project since 2003 when the Ministry of Water Resources, with joint efforts by the China Meteorology Administration and the Ministry of Land Resources to carry out nation-wide Sediment-related Disaster Prevention Guideline. The guideline mainly focuses on identification of disaster-prone area, establishment of a warning and evacuation system, relocation of existing houses in these areas. China has completed sediment disaster hazard map to identify the disaster-prone area and setting parameters of forecasting model to establish warning system. ¾ Producing sediment disaster hazard map Each Province made its hazard map on the basis of Sediment-related Disaster Prevention Guideline, including survey for conditions of precipitation, topography, economy and society. Following are some maps (Figure 3.4, 3.5, 3.6, 3.7) of Shanxin Province in west China.

31

Figure3.4 Rainfall Distribution Map Figure3.5 Topography Map

Figure3.6 Socio-economic Map Figure3.7 Sediment Disaster Hazard Map

¾ Setting parameters of forecasting model We use analytic method to do real time forecast of sediment disaster. The core of analytic method is setting the standard rainfall based on the series of observed data. Choosing continuous antecedent rainfall as the ordinate and 24-hour rainfall or other rainfall intensity as the abscissa, debris flow-causing rainfall and non-causing rainfall are plotted in the figure by different symbols respectively. Then those two rainfall groups are separated with a critical line or zone. If the rainfall exceeds the critical line or zone, a debris flow may be caused. Here takes Shijiagou basin, a tributary of the upstream of Yangtze River in Sichuan Province, a region frequently hit by sediment disasters because of its topographical and geological features, as an example to set parameters of forecast model. Table3.2 collects continuous antecedent rainfall and 24-hour rainfall from year 1959 to 1996, and classifies them into debris flow causing rainfall and non-causing rainfall respectively. Plot debris causing rainfall and non-causing rainfall in following figure by different symbols respectively (see Figure3.8). According to the figure, some

32 points can be achieved easily: z When the continuous antecedent rainfall is 66 mm or more and 24 hours rainfall is 28 mm or more, the debris flow may be caused. z When continuous antecedent rainfall is less than 66 mm and 24 hours rainfall is 62 mm or more, the debris flow may also be caused. ¾ Working Plan The detailed working plan for next step has also been worked out as listed below: verifying parameters of model and incorporating the model into operational forecast system, operational forecasting using the verified Model, Issuing Warning message if necessary.

Table 3.2 Debris Flow Causing rainfall and Non-causing rainfall in pilot area

dangerous

safe

Figure3.8 The X-Y graph to set parameters

33 „ ‘Development of guidelines for reservoir operation in relation flood forecasting’

China has constructed 86,000 reservoirs with combined storage capacity at 471.7 billion cubic meters; among which 422 belong to the large-scale reservoirs with multifunction of water supply and flood control. China is a country suffering both flood and water shortage due to unevenly distributed water resources. As a result how to rationally control floods for better managing water resources is one important topic for China’s social and economic development. For this purpose, developing reasonable guidelines for reservoir operation in relation flood forecasting is a very important issue in China. China has made great effort to study this issue and achieved some progress in dynamically regulating the reservoir.

„ ‘Evaluation and improvement of operational flood forecasting system focusing on model performance’ China has a vast territory with various climate features. From South to North, the country can be divided into equatorial zone, tropical zone, subtropical zone, warm temperate zone, temperate zone and cold zone. Different types of floods occur over two-third of the country’s land area every year. Therefore how to choose the suitable model to operationally forecast is very important, and deserves more attention. The main models in the flood forecasting system in China includes Xinanjing model, Shanbei(Excessive infiltration)model, API Model, Sacramento model, Unit hydrograph, Nash Unit Hydrograph Method, Maskingum routing method, Lag/K routing method, Liner diffusion wave routing method, Dynamic Wave Model, and some experiential methods. In South China and Northeast China, the humid regions with good vegetation conditions, models of runoff formation at the natural storage are suitable for flood forecasting, such as Xinanjiang model. In North China, the arid regions with poor vegetation conditions, models of runoff yield in excess of infiltration are suitable for flood forecasting, such as Shanbei model (like the Horton model). Generally in humid and some semi-humid regions the accuracy of flood forecasting proves to be more satisfactory.

2. Progress in Member’s Important, High-Priority Goals and Objectives a. Flood Forecasting System There have been new development for National Flood Forecasting System (NFFS) in year 2004. Some advanced technologies such as new hydraulic models which are more appropriate for river network regions and extended streamflow prediction（ESP）module have been successfully incorporated into NFFS. It means that NFFS is more powerful to forecast when facing various water structures such as

34 dams, gates, reservoirs and so on. And it also can provide long term runoff information for decision makers especially when water shortage becomes an urgent issue in China. b. Standard Development The new ‘Standard for Hydrological Code (SHIC)’ has been put into application in some pilot areas. Based on this Standard, the Standard for Structure and Identifier in Real-time Hydrological Information Database is under examination and approval, and it will be put into use in year 2005. To meet the development of automatic hydrological measurement and reporting system and data transmission technique, a new Data Transfer Protocol is under development. c. International Cooperation China is always keen on looking for international cooperation with neighboring countries and TC members, and taking an active part in the activities of international hydrological and meteorological organizations. We keep exchanging and sharing data with neighboring countries and organizations, such as Russia, Korea, Vietnam, India, Kazakstine and Mekong River Committee. In addition, a Chinese delegation which was led by Vice Minister of the Ministry of Water Resources of China visited D.P.R. Korea in September 2004 at the kind invitation of Hydrometeorological Bureau of D.P.R. Korea and a revised agreement on hydrological cooperation was signed.

3. Opportunities for Further Enhancement of Regional Cooperation More cooperation should be done among TC members in hydrological instruments and telecommunication equipment and flood forecasting system and in other areas of common interests. Every TC member is urged to further regional cooperation not only in the form of short term Workshop or TC Session alone, but also in conducting long-term training and expert exchanges.

35 IV. Disaster Prevention and Preparedness (DPP)

1. Progress in Member’s and Regional Cooperation and Selected RCPIP goals and Objectives a. Hardware and/or Software progress

Nil. b. Implication to Operational Progress Nil. c. Interaction with users, other Members, and/or other components On 25th May, the Ministry of Civil Affairs of China (MOCA) and the Secretariat for the UN International Strategy on Disaster Reduction (UN/ISDR) co-sponsored a 3-day International Conference on Disaster Reduction. The participants represented 18 countries including Bangladesh, Japan, Russia, Republic of Korea, and Lao People’s Democratic Republic and 7 international organizations including UNICIEF, over 20 national ministries and agencies, and more than 10 local departments of civil affaires in China. The conference focused on its discussion on the following issues: „ Disaster reduction assessment and disaster risk management; „ Improvement of capability and effectiveness of early warning; „ Disaster reduction and national sustainable development strategy as well as national public policy; „ Regional aspect for disaster reduction and cooperative mechanism and information sharing; „ Enhancing bi-lateral and multi-lateral emergency aid and human resources training; „ Recommendations on the preparation for the World Conference on Disaster Reduction. The conference identified the high-priority areas for the governments in Asia and adopted the Beijing Consensus that was to be submitted to the 2nd World Conference on Disaster Reduction in 2005 for discussion with an aim to guide the international disaster reduction practices. d. Training Progress

Nil. e. Research Progress

Nil. f. Other Cooperative/RCPIP Progress

Nil.

36 2. Progress in member’s Important, High-Priority Goals and Objectives a. Hardware and/or Software progress

Nil. b. Implication to Operational Progress Meteorological services at all levels have taken measures in improving their emergency management on meteorological disasters, setting up a direct reporting system and enhancing the assessment on high-impact meteorological-related disasters. Accordingly, the China Meteorological Administration (CMA) has made timely assessments on the impacts of typhoon Rananim and Aere. CMA has worked out Provisional Measures on the Issuance of Warning Messages for Contingent Meteorological Disasters on 16 August 2004. The measures have defined 11 types of warning messages, and relevant criteria and approach to issuing. These Measures will play a positive role in increasing the public awareness of disaster reduction and prevention and effectively preventing and mitigating the meteorological disasters. CMA has devoted itself to the development of the Government Contingency Program on Meteorological Disasters, a program for disaster management at national level in China, which clearly defines the deployment system, prevention and alert mechanism, emergency response, assurance and management. This Program is helpful to speed up the responding process and improve the emergency management through effective implementation of contingency program. During the period of Typhoon Rananim, the Ministry of Civil Affairs of China (MOCA) launched a 3-level disaster contingency program immediately after the disaster. Task groups representing civil affaires, finance, transportation, water, agriculture and meteorology made great effort in the post-disaster relief. c. Interaction with users, other Members, and/or other components

Nil. d. Training Progress

Nil. e. Research Progress

An atmospheric field experiment of tropical cyclone landfall with the acronym of CLATEX (China Landfalling Typhoon Experiment) was launched in July-August 2002. Some advanced instruments such as wind profiler, Doppler radar, ultrasonic anemometer, optical rain-gauge, radiosonde, satellite observation, tower observation, automatic weather stations and conventional upper level and surface observation etc. were utilized. Real-time intensive data from various sources of boundary layer and upper level of landing typhoons were acquired

37 and the target typhoon database including more than twenty categories of data was established.

The data (1901-2000) of centennial landfall tropical cyclones over China were processed and a data bank was established. The details of the data are including the best tracks, intensity, precipitation and wind distributions etc. Correspondingly, a centennial landfall typhoon service system via a network was set up to serve the public. f. Other Cooperative/RCPIP Progress

Nil.

3. Opportunities for Further enhancement of Regional Cooperation

Nil.

38 V. Typhoon that Impacted TC Members

1. Operational Forecast By September this year, six tropical cyclones (0407, 0409, 04xx, 0413, 0417 and 0420) landed over coastal areas of China. The table5.1 gives the mean distance errors of prediction of these tropical cyclones. It shows that the 24h, 48h and 72h mean errors of NMC are about 127, 227 and 332km respectively. These errors are very close to those of RSMC-Tokyo and JTWC. In the course of testing, we use real-time position of NMC as the basis of testing.

Table5.1 Mean distance errors of prediction of tropical cyclones

landing over China from Jan. to Sept. in 2004 (km) Forecast time 24h 48h 72h NMC/China 127 (101) 227 (73) 332 (58)

2. Narrative Accounts of Tropical Cyclones a. Characteristics of Landing Tropical Cyclones

As mentioned above, from Jan.1 to Sept.30 2004, 22 tropical storms in total were formed over the western North Pacific and the South China Sea. 6 make landfalls over China during 9 months (see table5.2).

Table5.2 List of Tropical cyclone landing over China from Jan. to Sept. in 2004 Maximum wind Minimum Number/Name Landing location Time/Date speed when SLP When landing (m/s) landing (hPa) Hualian, province 14:40UTC,Jul.1 30 980 Mindulle(0407) Yueqing, Zhejiang 01:30UTC,Jul.3 25 985 Kompasu(0409) 07:00UTC,Jul.16 23 990 Nameless(04xx) Huilai-haifeng, Guangdong 03:00UTC,Jul.27 20 995 Rananim(0413) Wenling, Zhejiang 12:00UTC,Aug.12 45 950 Fuqing, Fujian 08:30UTC,Aug.25 35 970

Aere(0417) Shishi, Fujian 13:30UTC,Aug.25 33 975 Longhai, Fujian 18:30UTC,Aug.25 30 980 Dongshan, Fujian 02:30UTC,Aug.26 20 995 Haima(0420) Wenzhou, Zhejiang 04:00UTC,Sept.13 18 998

Table5.2 shows that the intensity of 6 landed tropical cyclones was relatively weak when they landed. Only two of them were in typhoon category, one was severe tropical storm, and three of them were tropical storms.

39 b. Narrative on Tropical Cyclones

„ Mindulle (0407)

Tropical storm Mindulle (0407) was formed in the afternoon on June 23 over Western North Pacific, and then it moved westward in 15 kilometers per hour and became intensified enough to be a typhoon in the afternoon on June 27. On June 28, Mindulle slowed down and entered the Bashi channel and its intensity reached the climax on the next day. From June 29 to 30, after being detained in the Bashi channel for one day, Mindulle gradually turned to north and moved towards the coastal areas of southeast Taiwan province by going through Bashi channel. With its intensity greatly reduced by the terrain, the typhoon changed to the severe tropical storm and landed at Hualian, Taiwan province at 14:40UTC on July 1, on which the max winds of 30m/s near center. Mindulle afterwards continued to move northward, after passing through northern Taiwan province and entering the south of East China Sea, it landed again on Yueqing, Zhejiang province on 1:30UTC on July 3, with its max speed of 25m/s near center during the landing. After its second landing, Mindulle reduced to a tropical storm and moved north-northeast with a quicker pace. After going through coasting areas of east of Zhejiang province, it moved out of the mainland and reentered into the East China sea heading towards northeast near the Zhoushan, Zhejiang province in the evening of July 3. On the next day, this tropical storm quickly passed through south of Yellow sea and landed once again on the south of Korea. It continued to move northeastwards after that and faded away in the Korea peninsula (Figure5.1a, 5.1b). It could be seen from the above description, the typhoon Mindulle firstly had a long lifespan as it survived 282 hours. Next, the number of its landing is more than usual since it landed on three different places: Taiwan province, Zhejiang province and Korea peninsula. The third one to be noted was that the intensity of the typhoon underwent a tremendous change and its path is rather complex. 18 people died or injured wounded. The economic loss was estimated about 109 million yuan.

Figure5.1a Infrared imagery at 0030UTC on Jul. 3，2004 Figure5.1b Track of Typhoon Mindulle (0407)

40 „ Kompasu (0409)

Tropical storm Kompasu (0409) was generated in the morning on July 14 over Western North Pacific away from Taiwan province about 500km. Afterwards it moved westwards with a speed of 15-20 kilometers per hour and entered into the northeast of South China sea after passing through the Bashi Channel. At 7:00UTC July 16, Kompasu landed on the littoral of eastern Hong Kong, with the max winds of 23m/s near center. After landing, Kompasu moved northwards with its intensity reduced and faded away in Guangdong province (Figure5.2a, 5.2b). The lifespan of this tropical storm was relatively short as it only survived 60 hours and the intensity was weak with a max speed of 23m/s near center. The whole moving track takes on a shape of reversed ‘S’.

Figure5.2a Infrared imagery at 0630UTC on Jul. 16, 2004 Figure5.2b Track of Tropical Storm Kompasu (0409)

„ Nameless(04xx)

Tropical storm Nameless was developed in the morning of July 27 over northeast of South China Sea, which was only 35 kilometers away from the beach of Guangdong province. After its generation, it moved north-northwest，three hours later，it landed on the areas between Huilai and Haifeng of Guangdong province with the max winds of 20m/s near center. Then it reduced into a tropical depression (Figure5.3a, 5.3b). Having lasted for 30 hours, this tropical storm had a shorter lifespan than usual. 10 persons died or wounded. The economic loss was estimated about 31 million yuan.

Figure5.3a Infrared imagery at 0230UTC on Jul. 27, 2004 Figure5.3b Track of Tropical Storm Nameless(04xx))

41 „ Rananim (0413)

The 13th tropical storm Rananim was generated in shape evening on August 8, 2004 over Western North Pacific, east of Island of the Philippines. It then approached towards the littoral of southeast of China along the northwest path and turned into a severe tropical storm on August 10 and a typhoon the next day. Rananim began to move faster after it entered into the south of East China Sea in the evening of August 11, and landed at Wenling, Zhejiang province on the next night, with the max winds of 45m/s near center. Rananim turned to westwards after landing and passed through the central areas of Zhejiang with its intensity being weakened. In the wee hours of August 13, it was weakened into a severe tropical storm and then a tropical storm moving into the province. After lingering for 22 hours, Rananim moved to the northeast of province in the morning of August 14 and then to southeast of Hubei province. It was gradually reduced into a depression and disappeared (Figure5.4a, 5.4b, 5.4c, 5.4d) Typhoon Rananim was well known for its intensive wind, large quantity of precipitation and high tide. It attacked the Zhejiang province from the right side and brought about a great deal of damage. Rananim was the most intensive typhoon landing on Zhejiang province for 48 years since the landing of Typhoon Wanda in 1956 and the strongest typhoon landing on the mainland since Typhoon Sally (9616) in 1996 with 2263 casualties. The economic loss was estimated about 20.1 billion yuan.

Figure5.4a Wenzhou Radar echo at 1100UTC. Figure5.4b Wenzhou Radar radial velocity at on Aug. 12, 2004 0756UTC. on Aug. 12, 2004

Figure5.4c Infrared imagery at 1130UTC on Aug. 12, 2004 Figure5.4d Track of Typhoon Rananim (0413)

42 „ Aere (0417)

The 17th tropical storm Aere of this year was formed in the morning of August 20 over Western North Pacific. After moving northwestwards with a speed of 15-20 kilometers per hour, it grew into a severe tropical storm in the next morning and a typhoon at night. On August 24, Aere changed its course westwards east of and entered into the via north of Taiwan province. Then it suddenly quickened its pace and approached the littoral of central Fujian province. On the next day, Aere first landed on Fuqing, Fujian province at 8:30UTC with the max winds of 35m/s near center. Then it turned to move southwestwards and made landfall over Fujian province three times again. Finally it entered into Guangdong province along the southeast coastal line of Fujian province. On August 26, Aere was reduced into a depression and then disappeared in Guangdong province (Figure5.5a, 5.5b, 5.5c, and 5.5d). 11 persons died or injured. The economic loss was estimated about 3.2 billion yuan.

Figure5.5a Xiamen Radar echo at 0012UTC. Figure5.5b Changle Radar echo at 1132UTC on on Aug. 25, 2004 Aug. 25, 2004

Figure5.5c Infrared imagery at 0830UTC Figure5.5d Track of Typhoon Aere (0417) on Aug. 25, 2004

„ Haima (0420)

The 20th tropical storm Haima formed over sea northeast of Taiwan province at early morning on Sept. 12, 2004. Then it moved northwestward by north with a speed of 10-15 kilometers per hour and entered into the south of East China Sea after crossing coastal waters of northeast of Taiwan province. Afterwards Haima gradually turned to move

43 northwestward and finally landed over Wenzhou, Zhejiang province with the max winds of 18m/s near center at 04:00UTC September 13. After landing, Haima quickly weaken a tropical depression and disappeared in south of Zhejiang province (Figure5.6a, 5.6b).

Figure5.6a Infrared imagery at 0000UTC Figure5.6b Track of Tropical Storm on Sept. 13, 2004 Haima (0420)

VI. Resource Mobilization Activities

Nil.

44 VII．REPORT ON DAMAGE CAUSED BY CYCLONES, FLOODS AND DROUGHT

COUNTRY : China PERIOD COVERED BY THIS REPORT from : 1 Jan. 2004 to : 30 Sept. 2004 (date, month, year) (date, month, year) PREPARED AND SUBMITTED BY:

DATE PREPARED : 16 Oct. 2004 (date, month, year)

INTRODUCTION i．It was decided at the fourteenth session of the Typhoon Committee (Manila, November 1981) that information on damage caused by typhoons and floods should be compiled and sent to the Typhoon Committee Secretariat (TCS) before each annual session of the Typhoon Committee. This information shall consist of statistics on loss of human life, damage to houses, public facilities, agricultural products, etc. ii． At the fifth session of Management Board of the Typhoon Operational Experiment (TOPEX) (Tokyo, February 1982) UNDRO and LRCS were asked to co-operate in the preparation of a simple standard format for the region and make proposals for consideration by the Board at its sixth session. iii． The Board considered the proposed format at its sixth session (Bangkok, November 1982) and requested ESCAP and WMO in consultation with UNDRO and LRCS to revise the format with a view to incorporating more elaborately ESCAP long experience in flood statistics and to avoiding duplication with the ongoing efforts of ESCAP to improve disaster statistics. iv． Accordingly, this format was prepared for consideration at the third Planning Meeting for TOPEX (Tokyo, February 1993). The revised format was considered and adopted by the Meeting after some minor editorial amendments.

REPORT 1. This report should cover the total damage caused by typhoons and heavy rainfall, and associated storm-surges, floods, landslides, etc. 2. This report should be prepared by an official of the agency responsible for the disaster preparedness and relief in consultation with other agencies concerned. * Such official should be designated by each member and reported to TCS beforehand.

FORMAT 1. This format is designed to aid compilation of data and information which are already collected in each country. In other words, it does not propose any change in the existing systems of disaster damage survey in the various countries. 2. If final official figures for the reporting period are not available, it is recommended that tentative data be reported with appropriate notations. 3. Although this format covers broad aspects of disasters and detailed data, if the country is not prepared to provide data on some of the items, those may be left blank. However, it is recommended that the country report provides data at least on vital items marked with an asterisk and enclosed thick lines which are regarded as basic elements in disaster statistics on typhoon damage. 4. Data processing involved in the estimation of damage costs require much time, therefore, if the data are still being processed at the time of reporting, it should be noted when such data will become available.

* = Applicable for the members of Typhoon Committee.

45 Noted： For consistency, please use the following necessary: ... data are not available or not separately reported .. amount is negligible or nil N/A item is not applicable

I. GENERAL Sequence No. 1 2 3 4 5 6 7

1. Type of disasters 0407 0409 04xx 0413 0417 0420

Sequence number/code name of the typhoon and or type of MINDULLE KOMPASU Nameless RANANIM AERE HAIMA disaster caused by it or by a combination of weather disturbances such as rainfall, strong winds, storm-surges, floods and landslides. 1to3, JUL. 16, JUL. 27, JUL 12, AUG. 25, AUG. 13, SEPT. 2. Date or period of occurrence Zhejiang Guangdong Zhejiang Fujian Zhejiang Shanghai Fujian Zhejiang 3. Name of regions/areas seriously affected* Jiangsu Shanghai

II. HUMAN DAMAGE Unit

4. Dead and missing* persons 3 32 192 2

5. Injured persons 15 .. 7 2095 9 persons 6. Homeless* families persons 400 18182000 4401400 53000 7. Affected families 80 3636400 880280

8. Total persons

1) Please specify other categories of disaster victims covered here e.g. assisted by emergency relief, activities, those whose normal activities are seriously disrupted. Remarks:

46 III. MATERIAL DAMAGE IN PHYSICAL TERMS Sequence No. 1 2 3 4 5 6 7

A. Houses and buildings Unit

9. Destroyed* Units 290000 749770000 109050000

10. Damaged* Units 780000 2125250000 525060000

11. Affected* Units

12. Total* Units

B. Farmland

13. Farmland hectares 20130 740860 127900 7800

C. Agricultural Products

14. Crops tons

15. Livestock heads number 16. Fruit plants hectares

17. Others

2) Houses and buildings includes public buildings and are classified into three groups: Those not able to be used without reconstruction enter into destroyed”, those which can be required enter into damaged” and others which were inundated, damaged in minor parts or those fixtures and furniture were damaged enter into affected”.

3) Please specify other types of damage e.g. inundated marooned, evacuated.

4) Farmland affected are those buried, washed away, inundated and/or whose products were damaged.

5) If data are available for other products such as vegetables, marine products, forest products, please use this column.

Remarks:

47 Sequence No. 1 2 3 4 5 6 7

D. Public works facilities Unit

18. Road km 1163.1

19. Bridge sites

20. River embankment km 562.8 35.6 hectares 21. Irrigation facilities sites

22. Reservoir and dam number 380 1322 number 23. Harbour and port sites

24. Other please specify

B. Public Utilities km 25. Railway sites affected 26. Electric Supply families sites (km) 3342 affected 27. Water Supply families sites circuits 28. Telecommunication sites (km) 1522.1

29. Other please specify 6) There are two types of classification methods in the public works facilities: a) Classification in accordance with the nature of the service provided; b) Classification in accordance with the administrative structure of the government. Although the format was prepared according to the former classification, if necessary appropriate changes might be allowed. 7) Public utilities include both private owned and state owned facilities. Column of 揙 ther” can be used for the damage in airport, gas supply, etc.

48

Sequence No. 1 2 3 4 5 6 7

F. Others Unit

30. Ships lost or damaged number 18

31. Landslide and collapse of slope sites

IV. MATERIAL DAMAGE IN MONETARY TERMS Sequence 1 2 3 4 5 6 7

32. Damage of houses and loss of private property* includes: ∗ houses and buildings for residential use, ∗ household furniture, appliances and possession, ∗ stored good and other assets of farmers and fisherman 抯 households ∗ Other

33. Loss of agricultural production includes: ∗ crops, vegetables, fruits, etc. ten thousand 1365 6723 dollars ∗ livestocks ∗ Other: Fisheries 8) Damage of houses and loss of private property includes damage to a) houses and buildings for residential use; b) household furniture, appliances and possessions ; c) stored goods and other assets of farmers’ and fishermen 抯 households. Damage to shops and manufactures could be classified under item 34. Loss of industry, houwever, if such separation was not possible for small shops and home-industries, such damage could be included in this item with an appropriate note. Damage costs can be estimated by means of surveys listing the number of houses and buildings, their floor area and extend of damage, priced according to the value of the building or per unit area of floor space. Damage to household articles and personal effects such as clothing, furniture, electric appliances, cars, etc. are included in this category. If information on the household articles of an average family is available; loss may be calculated by multiplying the number of affected families by their total properties and an assessed percentage of damage. Damage to stored goods and other assets of farmers’ and fishermen 抯 household can be assessed in a similar manner. 9) Loss of agricultural production includes damage to a) crops, vegetables, fruits, etc., b) livestock, c) marine products, d) forest products. Damage to agricultural products which had been stored in farmers’ houses or warehouses should be counted under item 32. Damage of houses and loss of private properties. Crop damage can be estimated by multiplying the damaged crop area by the average loss per hectare and unit price of the crop, after considering the extent of damage to crops inundated and buried under debries. Loss of livestock can be estimated in the same manner by multiplying the head of stock lost by unit market price.

49

Sequence No. 1 2 3 4 5 6 7

34. Loss of industry ten thousand dollars

35. Loss of public work facilities includes items under III. ten thousand MATERIAL DAMAGE IN PHYSICAL TERMS dollars ∗ road bridge, river embankment, etc., irrigation facility ∗ reservoir and dam, harbour and port, and public bridges ∗ rehabilitation cost of farmland at government expense ∗ Other

36. Loss of public utilities includes items under III. MATERIAL DAMAGE IN PHYSICAL TERMS ∗ railway, electric supply, water supply, telecommunication ∗ Other

37. Total estimated/counted damage cost, sum of items 32, ten thousand 901 256 166541 17027 640 33, 34, 35, 36 dollars

10) Loss of industry includes damage to buildings, factories, warehouses, machinery, stored goodand other assets in factories and wholesale, retail and other service industries, but excludes agriculture, fishing and public utilities. Indirect losses due to suspension of routine activities are excluded here and if such data is available, please use column V. OTHER ADDITIONAL INFORMATION AND DATA AVAILABLE.

Estimates of the damage incurred can be sought from the industries concerned.

11) Loss of public works facilities is the cost required for the following facilities at Government expense: a) road and bridges, b) flood control installations, c) agricultural land, d) irrigation and drainage installations, e) reservoirs and dams, f) harbour, fishing port and airport installations, g) erosion control and landslide structures, h) streets, urban sewerage system and other public works facilites.

12) Public utilities include both private ownd and state owned facilites.

50

People’s Republic of China