Studi Baseflow Menggunakan Perbandingan 6 Metode RDF (Recursive Digital Filter)

Abstract

This research was conducted to estimate the baseflow at eight watersheds in UPT PSDA Bondowoso (East Java, Indonesia). Six method filter used for this study: (1) One Parameter Algorithm; (2) Boughton (Two Parameter Algorithm); (3) Ihacres (Three-Parameter Algorithm; (4) BFLOW (Lynie & Hollick Algorithm); (5) Chapman Algorithm; and (6) EWMA Filter. Methodology consist of : (1) data inventoring (2) data processing, (3) calibrating and validating, and (4) models performance evaluating. Daily discharge were used as main input for the analysis, then each recursive digital filter used to separate the baseflow from total flow. Furthermore, each algorithm was calibrate using daily discharge data for each year. The mean values of parameters obtained used to separate baseflow for whole period of record. RMSE, scatter plot, and FDC used to evaluate the model performance during dry period (Juli to September – with assumtion that there was no rainfall occurred in this period). Calibration was conducted on each watershed in the same technique. Validation was conducted from the watershed complete data (Kloposawit watershed) to others watersheds. This result showed that Lynie Hollick and EWMA perform better than others algorithms.