A new approach to use of wavelet transform for baseflow separation of Karst springs (case study: Gamasiyab spring)

Abstract

Today, groundwater is the essential resource the whole water sector demands. The spring baseflow discharge analysis can significantly help us better manage water resources. This study was focused on the separation baseflow discharge of the Gamasiyab karstic spring (western Iran) using the wavelet transform (WT). The recursive digital filter (RDF) method with different algorithms has been used to compare the approach performance. Forty-nine years (1968–2018) of the recorded springs’ daily discharge data were used for this aim. The isotopic content (IC) method was used to determine the best RDF algorithm and WT for baseflow separation. The IC of the monthly samples taken in 2017–2018 was measured. The results showed that the WT had the adjusted coefficient of determination (R_{{{text{adjusted}}}}^{2}) of 0.95 and root mean square error (RMSE) of 0.552 (m3/s) compared to the IC index method. The best algorithm in the RDF method was related to the Eckhardt algorithm with R_{{{text{adjusted}}}}^{2} of 0.97 and RMSE of 0.670 (m3/s), which had the best performance compared to the IC index method. Results showed that in most samples, the values of the separated baseflow by the WT are closer to the isotopic values. Also, based on the RMSE criterion, the WT has a lower RMSE than the RDF method and is known as the best method in this study. The results showed that the WT accurately separated the spring baseflow discharge and could split the baseflow of rivers.