Novel information for source identification of local pumping and recharging in a groundwater system
Hung-Jen Liu, Nien‐Sheng Hsu · Hydrological Sciences Journal · 2014
The aim of this paper is to solve the local pumping and recharging source identification problem by introducing a statistical technique, independent component analysis (ICA), into groundwater modelling. Such a problem suffers from limited observations in the field and is ill-posed. With the help of ICA, the observed head at monitoring wells is decomposed into a linear combination of head fluctuations caused by each source. The ill-posed problem is transferred into several independent well-posed ones. Moreover, the decomposed hourly head fluctuation, specially referred to as ‘characteristic pattern’, provides novel information on the type and distance of local sources. A hypothetical case to identify two local sources using two observation wells is presented, and the results show the accuracy and efficiency of the proposed method. The identified results are stable as long as the observation error is within a reasonable range. The characteristic pattern provides novel information, which can be widely used for source identification in the future. Editor D. Koutsoyiannis; Associate editor Xi Chen