In recent years, time series clustering is increasingly being applied in hydrogeology. Groundwater level fluctuations in long time series provide essential information to identify different hydrological behaviors and to validate the conceptual model of groundwater flow in aquifer systems. Piezometers also register the response to any changes that directly affect the amount of available groundwater resources.
On February 2019, the Guadalquivir Hydrographic Confederation (CHG) initiated the process to declare overexploited three of the five groundwater bodies of Almonte-Marismas aquifer, which host Doñana area. This new framework highlights the importance of studying historical groundwater level dynamics when defining new water resources management plans. In this work, groundwater level time series from 522 piezometers in the period 1975-2016 were analyzed. Data were preprocessed and transformed: selection of points, missing data imputation and data standardization. Cluster analysis was then applied to classify groundwater level hydrographs using available time series. Different groundwater levels fluctuations and responses were detected associated with the spatial distribution of geological properties in the aquifer. Moreover, actual hydrodynamics of the aquifer system were characterized making it easier for groundwater managers to make sustainable decisions.
Keywords: time series clustering, groundwater levels, hydrographs, water resources management.