Contents 1 Motivations and Objective 2 3 4 5 Radar and Flood Forecast System Study Area and Biseulsan Radar Hydrologic Analysis of Radar Rainfall Conclusions
Motivations and Objective In order to prevent flood disaster in a river basin, a flood warning system with real-time data collection system is required Installation and operation of fully-covered necessary rain gauge stations are not easy due to the budget and maintenance problem Radar rainfall estimates offer unique advantages - coverage over large areas - temporal updates as short as 10 minutes - high resolution in space MLTM installed rain radars to predict flood, so this study attempts to evaluate a hydrologic usefulness of the dual polarization radar rainfall estimation
Radar and Flood Forecasting in Korea Imjingang Radar Flood Control Office (FCO Data Collection System Modeling System: SFM < Characteristics of the major watershed > Sobaksan Radar FCC Item Area (km 2 % Raingage No. Stage No. Han R. 26,018 26.1 149 115 Bisuelsan Radar Nakdong R. 23,817 23.9 140 108 Keum R. 9,810 9.9 85 115 Seomjin R. 4,897 4.9 29 28 Youngsan R. 3,371 3.4 24 41 < Distribution of the major river and FCO in Korea > Total 63,016 68.2 427 407
Radar Rainfall Estimation Radar Rainfall Estimation Procedure Radar Observation Quality Control Generation of radar rain field Radar raw data (UF - All moments Removing AP, Clutter, etc Z, ZDR Bias Correction QCed radar data (UF -Qced_Z, ZDR, KDP Rain filed by LEMAP Rain filed by PPI Rain filed by CAPPI Radar Site DB Han River Flood Control Office standard DB Flood Control Office standard DB Ground gauge data Calculating point and basins Rainfall Adjustment by Gauge Estimation of radar rainfall Point rainfall Areal rainfall Minimum curvatures of R/G filed Rainfall analysis data R(<Z> R(<Z, ZDR> R(<Z, ZDR, KDP> Radar rainfall data
Radar Rainfall Conversion Algorithm 1. R(Z : Z-R relationship(traditional method R 2 0.714 ( Z = 1.70*10 * Z Z = 1.4 300R 2. R(ZDR : Bringi and Chandrasekar(2001 R( Z, Z DR 3 0.714 = 6.70*10 * Z ZDR 3.43 3. R(KDP : Rhyzhkov(2005 (NEXRAD prototype R( Z = 1.70*10 2 * Z 0.714 Being used in MLTM R( K DP = 44.0 K DP 0.822 sign ( K DP R = R ( Z / f ( Z if R ( Z < 6 mm / 1 DR h R = R( K DP / f 2 ( Z DR if 6 < R( Z < 50 mm / h R = R( K DP if R( Z > 50 mm / h where, f 1 (Z DR = 0.4 + 5.0 Z dr 1 1.3 Z = dr 10 0.1Z DR (db f 2 (Z DR = 0.4 + 3.5 Z dr 1 1.7
SFM Hydrologic Model Storage Function Method Model Simple computation procedure Considering nonlinear property for streamflow computation Inherent parameter estimation of watershed is very difficult < Schematic diagram of SFM model computation >
Flood Forecasting system Integrated Flood Forecasting System Channel Subbas Channel in - Routing Flood Forecasting System for urban area Flash Flood Forecasting System using rainfall radar
Study Area Namgang Dam Upstream Area Characteristics of Namgang dam basin Location App. 128 E, 35 N Basin Area 2,293 km 2 Basin Length 108 km Elevation 45 ~ 1,915 EL.m Land Use Mountain Annual MAP 1,514 mm Available data Raingauge St. 25 Stage St. 10 Dam 1 Sub-basins 14 Selected events : July ~ Sep. in 2012 * AE = Aneui, HY = Hamyang, MC = Macheon, IC = Imcheon SC = Sancheon, DS = Danseong, SG = Samga, SA = Shinan TS = Taeseong, CC = Changchon, NGD = Namgang dam < Study area of Namgang dam site >
Bisuelsan Radar Specifications Sub-System Components Specification Note Transmitter Antenna Receiver & RVP Data Frequency 2,795 MHz S Band Polarization H+V, H Dual Polarization Tube Peak Power Klystron 750KW Dish Size 8.5m Gain 45dB Beam width 0.95 Scan Speed 1 ~ 36 /sec Max 6rpm Scan Range AZ 0~360, EL -2 ~92 Receiver Type Heterodyne, Digital Dual Channels( H and V MDS Dynamic range -114dBm 100dB Signal Processor Digital I/Q Clutter Suppression Moments Time series Z, V, W, SQI, ZDR, PhiDP, KDP, RhoHV, Hclass Raw I/Q data Dual pol parameters
Selected Rainfall Event < Sancheong rainfall & stage station > TY Bolaven (27-29, Aug. TY Khanun (18-21, July Cont. Rain (22-26, Aug. TY Sanba (16-18, Sep.
Hydrologic Analysis of Radar Rainfall Comparison of Spatial Distribution (a Event 1(18~21, July (b Event 2(22~26, Aug. (c Event 3(27~29, Aug. (d Event 4(16~18, Sep. TM MAP TM MAP TM MAP TM MAP Radar MAP Radar MAP Radar MAP Radar MAP
Mean Areal Precipitation Difference of Radar Rainfall (a Event 1(18~21, July (b Event 2(22~26, Aug. (c Event 3(27~29, Aug. (d Event 4(16~18, Sep.
Characteristic Locality by Normalized Difference (a Event 1(18~21, July (b Event 2(22~26, Aug. (c Event 3(27~29, Aug. (d Event 4(16~18, Sep.
Temporal Variation of Radar Rainfall on SC Upstream Sancheong : 18~21, July Rainfall(mm/10m Radar Gauge Sancheong : 22~26, Aug. Rainfall(mm/10m Radar Gauge
Summary of Gauge-Radar Comparison Event 18~21, July 22~26, Aug. 27~29, Aug. 16~18, Sep. Number of gauges 24 24 24 24 Event duration(hours 48 69 46 49 Ave. G 74.63 211.79 140.33 266.40 Ave. R 61.87 198.65 122.60 237.13 Ave. difference(% 18.07 6.52 11.77 11.05 Rainfall cause Typhoon Frontal System Typhoon Typhoon
Hydrologic Application Radar-Streamflow Simulation on SC Upstream (a Event 1(18~21, July (b Event 2(22~26, Aug. (c Event 3(27~29, Aug. (d Event 4(16~18, Sep.
Radar-Streamflow Simulation on NGD Upstream (a Event 1(18~21, July (b Event 2(22~26, Aug. (c Event 3(27~29, Aug. (d Event 4(16~18, Sep.
Hydrologic Adjustment Trial Adjust radar rainfall estimation by correction factor Correction factor = 1.10 Correction factor = 1.06 (a Event 1(18~21, July (b Event 2(22~26, Aug. Correction factor = 1.10 Correction factor = 1.15 (c Event 3(27~29, Aug. (d Event 4(16~18, Sep.
Small and Upstream Area Performance IC Upstream A = 432km 2 Radar Diff. = 7.0 % Rada ar Gauge SA Upstream A = 413km 2 Radar Diff. = 6.2 % Radar Gauge
Conclusions and Remarks We have examined the hydrological application of the MLTM s dual polarization radar rainfall estimation The dual polarization radar-driven driven mean areal precipitation show a little differences compared with gauge observed rainfall as range as 6.2~18.1% underestimation The coupled radar-streamflow simulation results were also underestimated like as radar rainfall, However, the hydrologic adjustment trial showed a positive possibility to forecast flood by using the radar rainfall in actual operation The case study in the small and upstream watershed, the radar-streamflow simulation results had good performance, it would be useful to predict a flash flood in mountain area and a urban flood