GEOSCAN Search Results: Fastlink


TitleTying geophysics to hydrogeology: a learning machine approach to characterise heterogeneous granular aquifers
AuthorParadis, DORCID logo; Gloaguen, E; Lefebvre, R; Rivera, AORCID logo
SourceProceedings of GeoHydro 2011; 2011 p. 1-8
LinksOnline - En ligne (Full program/Programme complet, PDF 150 MB)
Alt SeriesEarth Sciences Sector, Contribution Series 20110244
MeetingGeoHydro 2011; Québec; CA; August 28-31, 2011
Subjectshydrogeology; geophysics; aquifers; groundwater; hydraulics; hydraulic analyses; hydraulic conductivity
Illustrationsflow charts
ProgramGroundwater Geoscience Aquifer Assessment & support to mapping
AbstractA learning machine approach is proposed to define site-specific hydro-geophysical relationships in order to predict granular aquifer hydraulic properties from geophysical measurements. The learning machine is trained on a representative data set of hydraulic and geophysical measurements. The main algorithms used for training are semisupervised fuzzy clustering and relevant vector machines (RVM) for classification and regression. This approach, which extends the capabilities of geophysical methods, represents an efficient alternative to conventional granular aquifer characterization mainly based on hydraulic methods.

Date modified: