Rain Forecasting for Ho Chi Minh City Using Doppler Weather Radar Dwsr-2500C

Abstract Rainfall amounts vary randomly over time and space. Rainfall monitoring and forecasting is a difficult task, especially for a short-term period from 30 minutes to 3 hours. Recently Doppler weather radars have been used as one of the new solutions in the short-term forecasting of extreme rain or storm. This research presents some results of forecasting the wind direction, velocity, and rainfall of a typical rainy day, 14 September 2010, based on CAPPI images of a DWSR-2500C radar in the Nha Be district, Ho Chi Minh City (HCMC). The results showed that the Doppler radar, in a scanning radius of 30 km, is very effective in forecasting extreme rainfall for each region and district when reflected radar signals from clouds moving towards the city are detected. This research provides useful information in the forecast of extreme rainfall for flood prevention works in the HCM City.


INTRODUCTION
HCMC is located between 10 0 10'N -10 0 38'N and 106 0 02'E -106 0 54'E in the south of Vietnam. It is a delta city and the biggest city in Vietnam. HCMC is vulnerable to flooding as a result of land subsidence, high rainfall and climate changes. The land area of HCMC is 2095 km 2 , and the average population density was 3719 persons per km 2 in 2012. The current city's population is expected to grow to 9.2 million by 2020 and 10 million by 2025 [1]. The rainy season is from May to November. Rainfall in the rainy season is approximately 80-85% of the annual rainfall. Heavy rains occur in June and September in the range of 250-330 mm/month. The maximum is up to 683 mm. The rainfall intensity is quite high -from 0.8 to 1.5 mm/minute [2]. Average annual rainfall is about 2,000 mm. However, the rainfall is not distributed evenly, but tends to increase in the southwestnortheast axis. Most central districts and northern districts usually have higher rainfall than districts in the south and southwest. Generally, the HCMC's rainfall is highly distributed in a short period of the year and causes water shortages and floods [3]. Therefore, rain forecasts for HCMC are very essential. Using a weather radar in rain forecasts is one of the most advanced techniques with the following advantages: ability to forecast extreme weather phenomena in a short period of time (30 minutes -3 hours), not achieved by other forecasting methods; highly accurate forecasting for the region near the radar within a radius of 120-240 km. The weather radar has been used to measure rainfall over hilly terrain in north Wales, England [4]. Ground-based radars offer basic and unique advantages, obtaining the continuous three-dimensional space scanning of precipitation events, a short volume-scan period, long range coverage and a high space resolution of measurements [5]. Radar analyses of extreme rainfall were described to clarify the flood-producing rainfall in Charlotte, North Carolina metropolitan area, USA [6]. Radar-rainfall products are crucial for input to runoff and flood prediction models, the validation of satellite remote sensing algorithms, and for the statistical characterization of extreme rainfall frequency [7]. Comparisons of radar and gauge measurements of rainfall were also conducted to examine the accuracy. There is a good agreement between rain gauges and radar rainfall estimates [8][9][10][11][12][13][14][15]. Using weather radars in the assessment of flash flood is very effective [16][17][18][19]. This paper shows the overview and results of applying the Doppler weather radar DWSR-2500C in extreme rain forecasts for HCMC.

Scanning radii of radar
The Doppler weather radar DWSR-2500C in the NB district uses the EDGE software system to manage and process information throughout its operations. The main function of EDGE consists in controlling the transmitter, receiver, antenna, signal processor of the radar; collecting information; arranging control functions, storage, data mining; creating products; transmitting data. Initially, the scanning radius of 240 km is used to get an overview of weather patterns. This is helpful to detect extreme weather phenomena from far distances in time. Afterwards, the scanning radii of 120 km, 60 km, or 30 km will be used to forecast more accurately.

Determining moving direction of rain
When detecting the presence of potential clouds through radar images, we need to concern and analyse signals reflected from the clouds to check whether they move into the city or not. These reflected signals are recorded during a certain period of time. It is also possible to use TRACK products of the radar to determine the moving orbit of the potential clouds.

Velocity of signals reflected from potential clouds
The velocity of the front boundary [20][21][22][23] of signals reflected from potential clouds to the interested location A (time from T n-1 to T n ) is calculated by the following formula (1): (n=0, 1, 2, 3...) (1) The time interval of the front boundary of the reflected signals from the location N to the location A is given by (2):

Figure 1: HCMC map and radar station in NB district
Rain starting time is the time when the reflected signal zone meets the location A and is expressed by (3): L n represents the distance from the n th boundary in front of the reflected signal zone to the location A (km).

Rainfall analysis from radar images -a typical case of 14 September 2010
Ground rainfall data was measured by a rain gauge in the Tan Son Hoa meteorological station, on 14 September 2010, in HCMC. It began to rain at 16h00 (9h00Z). To observe the overall situation, formation, and the movement of potential clouds which may cause rain in HCMC, we needed to analyse the radar images about 2 hours before the starting time of rain. The radar uses International time (Table 1).
For comparison with ground rainfall in the Tan Son Hoa station, we used CAPPI image products of the radar.

RESULTS
The results of analyses are shown below: + At 8:00Z: There is a cloud appearing in 10045'N -106037'E with a reflected intensity of 39.5 dBz and a rainfall rate of 10.7 mm/h. This is a local convective cloud causing rain in the TB and TP districts (Figs. 2&3).           (Fig. 15).   There are some places with an intensity of 44 dBz and a rainfall rate of 20.5 mm/h. The reflected signal zone has a tendency to move towards the south and causes rain in the western ridge areas; however, it does not move to the city centre (Fig. 19). Obviously, when comparing the rain event in Tan Son Hoa station (starting time 9:00Z) to the potential cloud moving towards the city detected by the radar at 8:00Z, we see that radar is able to forecast rain in a short time effectively.

CONCLUSION
Using the Doppler weather radar DWSR-2500C to monitor the change of rain and extreme weather patterns over wide areas gives fairly accurate results. Depending on the region and case, analyses require appropriate scanning radii. For rain forecasting in HCMC, the weather radar in the NB district initially uses the scanning radius of 240 km to observe weather patterns. Afterwards, the scanning radii of 120 km, 60 km, or 30 km will be used to forecast more accurately. It is realized that when the signal reflected from clouds moves closer to the city, the scanning radius of 30 km is very effective in a short period of time (30 minutes to 3 hours). This research provides useful information for flood prevention works in HCMC.