The analysis of the dynamic response of a deep karst aquifer to precipitation is a complex problem, due to the structure of the aquifer, which may origin non-linear variations of the groundwater table. However, karst aquifers may be one of the main water resources for those regions characterized by poor presence of shallow water and medium-high permeable soils. This work presents an analysis of the dynamic response of the large aquifer of central Apulia, south Italy, based on a data-driven approach, namely Evolutionary Polynomial Regression. Four wells were monitored, and for each of them about 15 years of monthly average levels are available. In particular, the dynamic response of the aquifer is modelled in terms of prediction of the groundwater levels as function of total monthly precipitations and past measured groundwater levels. A single model, as closed-form equation, is obtained for each well and then these models are compared, highlighting the differences and the similarities among the responses of each well.
Analysis of the dynamic response of a large deep karst aquifer by data-driven modelling / Doglioni, Angelo; Simeone, Vincenzo. - ELETTRONICO. - (2014), pp. 256-263. (Intervento presentato al convegno 22nd UK Conference of the Association for Computational Mechanics in Engineering tenutosi a Exeter, UK nel April 2-4, 2014).
Analysis of the dynamic response of a large deep karst aquifer by data-driven modelling
Angelo Doglioni;Vincenzo Simeone
2014-01-01
Abstract
The analysis of the dynamic response of a deep karst aquifer to precipitation is a complex problem, due to the structure of the aquifer, which may origin non-linear variations of the groundwater table. However, karst aquifers may be one of the main water resources for those regions characterized by poor presence of shallow water and medium-high permeable soils. This work presents an analysis of the dynamic response of the large aquifer of central Apulia, south Italy, based on a data-driven approach, namely Evolutionary Polynomial Regression. Four wells were monitored, and for each of them about 15 years of monthly average levels are available. In particular, the dynamic response of the aquifer is modelled in terms of prediction of the groundwater levels as function of total monthly precipitations and past measured groundwater levels. A single model, as closed-form equation, is obtained for each well and then these models are compared, highlighting the differences and the similarities among the responses of each well.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
