Aquifers are an essential source of water supply for human consumption, agricul tural irrigation, and industrial purposes. Understanding their complexity is crucial for sustainable groundwater management. However, estimating the properties of the aquifer and their spatial variability can be challenging, but it is necessary to accurately simulate groundwater flow and transport. In this study, an inverse ap proach is applied to infer the characteristics of a binary field using concentration data obtained from a tracer test carried out in a laboratory sandbox. The experi mental device reproduces a vertical section of an unconfined aquifer. Glass beads of two different diameters are used to mimic the binary porous medium, while flu orescein sodium salt served as a tracer. Breakthrough curves recorded at different locations are used as observations to estimate the spatial distribution of two litho types and their main hydraulic and transport parameters. The inverse problem is solved by coupling the ensemble smoother with multiple data assimilation (ES MDA), which is an ensemble Kalman-based method, with a truncated Gaussian model (ES-MDA-T). The advantage of the ES-MDA-T, with respect to the stan dard ES-MDA, is the possibility to simultaneously estimate different properties of the geological lythotypes and their spatial distribution with an acceptable amount of computational time. The proposed technique is tested using two methods: a fully parameterized scheme and a pilot point approach. In order to evaluate the capability of the methodology, a synthetic case that reproduces the sandbox ex periment was initially developed. Once the procedure was validated, the optimal configuration obtained from the synthetic case was used for the experimental case study. The results show that the ES-MDA-T method replicates well the binary field and its aquifer properties even in the presence of considerable measurement errors. The pilot point approach takes less computation time than the fully parameterized scheme, while yielding comparable results.
Coupling an ensemble smoother with a truncated Gaussian model for aquifer characterization / Todaro, V.; D'Oria, M.; Zanini, A.; Gómez-Hernández, J. J.; Tanda, M. G.. - (2023). (Intervento presentato al convegno The 22th Annual Conference of the International Association for Mathematical Geosciences tenutosi a Trondheim, Norway nel 5-11 Agosto 2023).
Coupling an ensemble smoother with a truncated Gaussian model for aquifer characterization
Todaro V.
;D'Oria M.;Zanini A.;Tanda M. G.
2023-01-01
Abstract
Aquifers are an essential source of water supply for human consumption, agricul tural irrigation, and industrial purposes. Understanding their complexity is crucial for sustainable groundwater management. However, estimating the properties of the aquifer and their spatial variability can be challenging, but it is necessary to accurately simulate groundwater flow and transport. In this study, an inverse ap proach is applied to infer the characteristics of a binary field using concentration data obtained from a tracer test carried out in a laboratory sandbox. The experi mental device reproduces a vertical section of an unconfined aquifer. Glass beads of two different diameters are used to mimic the binary porous medium, while flu orescein sodium salt served as a tracer. Breakthrough curves recorded at different locations are used as observations to estimate the spatial distribution of two litho types and their main hydraulic and transport parameters. The inverse problem is solved by coupling the ensemble smoother with multiple data assimilation (ES MDA), which is an ensemble Kalman-based method, with a truncated Gaussian model (ES-MDA-T). The advantage of the ES-MDA-T, with respect to the stan dard ES-MDA, is the possibility to simultaneously estimate different properties of the geological lythotypes and their spatial distribution with an acceptable amount of computational time. The proposed technique is tested using two methods: a fully parameterized scheme and a pilot point approach. In order to evaluate the capability of the methodology, a synthetic case that reproduces the sandbox ex periment was initially developed. Once the procedure was validated, the optimal configuration obtained from the synthetic case was used for the experimental case study. The results show that the ES-MDA-T method replicates well the binary field and its aquifer properties even in the presence of considerable measurement errors. The pilot point approach takes less computation time than the fully parameterized scheme, while yielding comparable results.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
