Author(s): Ratih Indri Hapsari; Satoru Oishi; Magfira Syarifuddin; Winda Harsanti; Fahmi Hidayat
Linked Author(s): Satoru Oishi, Fahmi Hidayat, Ratih Indri Hapsari
Keywords: Rainfall; Radar; Rain gauge; Variability; Rainfall-runoff model
Abstract: X-band weather radar has considerable potential for its fine spatio-temporal resolution which is significant in the hydrological simulation. In this study, we evaluate the distribution variability of precipitation observation from X-band radar and network rain gauges in Upper Konto River Basin Indonesia which is originated from Kelud volcano. The rainfall spatial representativeness and the temporal variability of X-band radar vs. nine rain stations are compared during the observation campaign in the rainy season of February 2017. The evaluation of spatial distribution characteristic uses Kagan interpolation method in minute scale, hourly, and daily total volume of precipitation. In order to study the impact on the catchment size, the outlet of the basin in Selorejo Reservoir (235.58 km²) is used. Performance of the BTOPMC physically distributed hydrological model using 150 m resolution radar-rain was evaluated over that of using lower resolution from rain gauges. The analysis reveals that widely spread gauges give lower spatial correlation and larger error for the smaller scale of time. The advantages of X-band observation for detecting the variation of rainfall in the small area leads to better reproducibility of flood simulation, shown by RMSE of 9.85 and 7.78 m³/s for rain gauge and radar respectively. It means the radar-rainfall input could give better reproducibility of runoff simulation. It is proven that the radar provides better representativeness of rainfall spatial and temporal variation for operational runoff simulation, which is important for daily as well as seasonal reservoir operations.
Year: 2018