THE TRMM RAINFALL-RUNOFF TRANSFORMATION MODEL USING GR4J AS A PREDICTION OF THE TUGU DAM RESERVOIR INFLOW
Keywords:Rainfall-runoff model, GR4J, Tugu dam, Nash-sutcliffe, TRMM
This study presents numerical modeling to simulate the rainfall-runoff model using Génie Rural à 4 paramètres Journalier (GR4J) of the Keser Watershed (East Java, Indonesia) to predict the reservoir inflow of the Tugu dam. The four independent parameters used to optimize the daily rainfall-runoff model are the maximum capacity of the production store (X1), the underground water exchange coefficient for the catchment (X2), the one-day capacity of the routing storage (X3), and the time base of the unit hydrograph (X4). The TRMM daily precipitation satellite time series is used as the rainfall input data spanning two years (2017-2018). Furthermore, the potential evapotranspiration for the catchment was calculated using the Penman method. The rainfall-runoff transformation results of the discharge simulation were then compared with observed discharge data collected from the hydrometry gauge station. The simulation results showed that the best performances of the model obtain 7.85, 2.52, 3.93, and -1.35 for the coefficient result of the maximum capacity of the production store (X1), the underground water exchange coefficient (X2), the one-day capacity of the routing storage (X3), and time base of the unit hydrograph (X4) respectively. In addition, the Nash-Sutcliffe coefficient's deviation value obtains 0.73, which is considered a good performance model (0.65<NSE<0.75). The percent bias (PBIAS) calculation obtains the deviation of 3.5 %, showing this GR4J model in Keser watersheds can be accepted (under 5%). Therefore, the coefficient of determination (R-squared) obtains 0.66, a good fit performance model for daily data and 0.951 for monthly data inferring that the model has a strong correlation performance.