The Reliability of W-flow Run-off-Rainfall Model in Predicting Rainfall to the Discharge

This research intends to predict the discharge (run-off) from rainfall for which the model is built using W-flow. The research location is in the Gajah Mungkur reservoir (Wonogiri) in Indonesia. The estimation of reservoir inflow has an important role, mainly in the scheme of reservoir operation and management. However, the heterogeneity of complex spatial and temporal patterns of rainfall and also the physiographic context of a watershed cause the development of a model of real-time run-off and rainfall that can accurately predict the reservoir inflow to become a challenge in the development of water resources. In relation to the analysis and prediction of rainfall, the constraint and problem that is still often faced is the minimal availability of observed rainfall data spatially as well as temporally; the time series of rainfall data is not long and complete enough; and the number of rainfall stations is less evenly distributed. The methodology consists of carrying out the literature study, collecting as much rainfall data as possible to build a W flow model, then carrying out the model calibration and analyzing the prediction of real-time reservoir inflow for operation. The result shows that the dependable discharge of the Wonogiri watershed shows that there are two peak discharges, which happened on February II (the second half of February) and December II (the second half of December). However, the discharge is decreasing in July and reaching its lowest level in October II (the second half of October).