Comparing automatic baseflow separation filters on three watersheds in the urbanised area of East Java, Indonesia

• Published in: Journal of water and land development 2021-11, p.130-140
• Main Authors: Indarto, Indarto, Hardiansyah, Mujiono, Sujarwo, Mohamad Wawan
• Format: Article
• Language: eng ; pol
• ISSN: 2083-4535; 2083-4535
• DOI: 10.24425/jwld.2021.138168

Baseflow is the primary source of water for irrigation and other water needs during prolonged dry periods; accurate and rapid estimation of baseflow is therefore crucial for water resource allocation. This research aims to estimate baseflow contribution during dry periods in three small watersheds in East Java: Surabaya-Perning (114 km2), Lamong-Simoanggrok (235 km2), and Bangsal-Kedunguneng (26 km2). Six recursive digital filters (RDFs) algorithms are explored using a procedure consisting of calibration, validation, evaluation and interpretation. In this study, the period of July to September is considered as the peak of the dry season. Moreover, data for the period 1996 to 2005 is used to calibrate the algorithms. By yearly averaging, values are obtained for the parameters and then used to test performance during the validation period from 2006 to 2015. Statistical analysis, flow duration curves and hydrographs are used to evaluate and compare the performance of each algorithm. The results show that all the filters explored can be applied to estimate baseflow in the region. However, the Lyne–Hollick (with RMSE = 0.022, 0.125, 0.010 and R2 = 0.951, 0.968, 0.712) and exponentially weighted moving average or EWMA (with RMSE = 0.022, 0.124, 0.009 and R2 = 0.957, 0.968, 0.891) for the three sub-watersheds versions give the best performance.