A novel integrated workflow for deep geothermal reservoir characterization using core data, optimized coupled petrophysical–seismic modeling, and machine learning: A case study from the SW Paris Basin, France
Abstract
The Paris Basin hosts one of the longest-operating deep geothermal systems in Europe, with approximately 50 geothermal doublets exploiting the Middle Jurassic Dogger carbonate and Lower Cretaceous Albian siliciclastic aquifers and supplying nearly 2 TWh of renewable heat annually. Despite extensive geological and geophysical investigations, predicting reservoir facies, heterogeneity, and transmissivity remains highly uncertain. This study presents a physics-guided, multi-scale workflow that integrates petrophysics, rock physics, seismic inversion, and stochastic reservoir modeling to characterize the Dogger, Oxfordian, and Triassic geothermal reservoirs of the Paris Basin. The workflow combines four components: (i) estimation of mineral volumes and total porosity ( ϕ total ) using a hybrid statistical–machine learning approach calibrated with X-ray diffraction (XRD) measurements; (ii) joint Monte Carlo inversion of water saturation (S w ) and formation water resistivity (R w ) under petrophysical constraints; (iii) generation of a rock-physics-guided synthetic pseudo-well catalogue and corresponding pre-stack Zoeppritz angle gathers, in which XRD-calibrated mineral end-members and petrophysically consistent elastic trends impose physical structure absent from a purely data-driven training set; and (iv) convolutional neural network (CNN) prediction of clay volume (V clay ), total porosity, and effective porosity (ϕ eff ) from pre-stack seismic data, pretrained on this synthetic catalogue and fine-tuned via transfer learning on seven real wells. Three additional wells, excluded from transfer learning, were reserved for semi-blind validation. At the calibration wells, R 2 ranges from 0.858 to 0.985 across the three properties, with normalized mean-square error (NMSE) of 1.8%–18.0% and root-mean-square error (RMSE) below 0.043. At the semi-blind wells, performance varies with local well-to-seismic tie quality: one well reproduces calibration-level statistics (R 2 > 0.92 , NMSE <10%), while the other two show progressively degraded performance (R 2 as low as 0.41, NMSE up to 113%), consistent with documented tie-quality issues at those locations rather than a general failure to generalize. The predicted properties reproduce major depositional trends, including Oxfordian lateral continuity and Dogger reservoir compartmentalization, without explicit stratigraphic conditioning. Effective porosity predicted along 1,990 km of interpreted 2D seismic profiles was then integrated into stochastic three-dimensional reservoir modeling, generating 100 equiprobable realizations that represent reservoir heterogeneity and uncertainty while remaining consistent with well observations and regional seismic trends
Related
- Climate change impacts on the predicted performance of optimized building-integrated photovoltaics
- Relationship between soil electrical conductivity and water content: A multiscale petrophysical calibration
- A flexible regularization for magnetic vector inversion with petrophysical, structural, and sparsity constraints
- Diatom-based treatment of geothermal brines to selectively remove silicon, retain lithium for downstream recovery, and produce omega-3 fatty acids
- A multivariate integrated drought characterization framework using GNSS-GRACE terrestrial water storage and precipitation
BibTeX
@article{pwavodi2026novel,
title = {{A novel integrated workflow for deep geothermal reservoir characterization using core data, optimized coupled petrophysical–seismic modeling, and machine learning: A case study from the SW Paris Basin, France}},
author = {Joshua Pwavodi and Camille Maurel and Aurelien Bordenave and Alexandre Stopin and Alexandre Ortiz and Agathe Faure and Eric Lasseur and Laurent Beccaletto and Justine Briais and Benoit Issautier and Stephane Marc and Nicolas Maubec and Laure Capar and Pavel Didenko and Kirill Siraev and Shervin Rasoulzadeh and Amit Singh and Ali Moradi Tehrani},
journal = {Geothermics},
year = {2026},
month = oct,
doi = {10.1016/j.geothermics.2026.103880},
url = {https://brgm.hal.science/hal-05845740v1},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - A novel integrated workflow for deep geothermal reservoir characterization using core data, optimized coupled petrophysical–seismic modeling, and machine learning: A case study from the SW Paris Basin, France
AU - Joshua Pwavodi
AU - Camille Maurel
AU - Aurelien Bordenave
AU - Alexandre Stopin
AU - Alexandre Ortiz
AU - Agathe Faure
AU - Eric Lasseur
AU - Laurent Beccaletto
AU - Justine Briais
AU - Benoit Issautier
AU - Stephane Marc
AU - Nicolas Maubec
AU - Laure Capar
AU - Pavel Didenko
AU - Kirill Siraev
AU - Shervin Rasoulzadeh
AU - Amit Singh
AU - Ali Moradi Tehrani
PY - 2026
DA - 2026/10/30
JO - Geothermics
DO - 10.1016/j.geothermics.2026.103880
UR - https://brgm.hal.science/hal-05845740v1
AB - The Paris Basin hosts one of the longest-operating deep geothermal systems in Europe, with approximately 50 geothermal doublets exploiting the Middle Jurassic Dogger carbonate and Lower Cretaceous Albian siliciclastic aquifers and supplying nearly 2 TWh of renewable heat annually. Despite extensive geological and geophysical investigations, predicting reservoir facies, heterogeneity, and transmissivity remains highly uncertain. This study presents a physics-guided, multi-scale workflow that integrates petrophysics, rock physics, seismic inversion, and stochastic reservoir modeling to characterize the Dogger, Oxfordian, and Triassic geothermal reservoirs of the Paris Basin. The workflow combines four components: (i) estimation of mineral volumes and total porosity ( ϕ total ) using a hybrid statistical–machine learning approach calibrated with X-ray diffraction (XRD) measurements; (ii) joint Monte Carlo inversion of water saturation (S w ) and formation water resistivity (R w ) under petrophysical constraints; (iii) generation of a rock-physics-guided synthetic pseudo-well catalogue and corresponding pre-stack Zoeppritz angle gathers, in which XRD-calibrated mineral end-members and petrophysically consistent elastic trends impose physical structure absent from a purely data-driven training set; and (iv) convolutional neural network (CNN) prediction of clay volume (V clay ), total porosity, and effective porosity (ϕ eff ) from pre-stack seismic data, pretrained on this synthetic catalogue and fine-tuned via transfer learning on seven real wells. Three additional wells, excluded from transfer learning, were reserved for semi-blind validation. At the calibration wells, R 2 ranges from 0.858 to 0.985 across the three properties, with normalized mean-square error (NMSE) of 1.8%–18.0% and root-mean-square error (RMSE) below 0.043. At the semi-blind wells, performance varies with local well-to-seismic tie quality: one well reproduces calibration-level statistics (R 2 > 0.92 , NMSE <10%), while the other two show progressively degraded performance (R 2 as low as 0.41, NMSE up to 113%), consistent with documented tie-quality issues at those locations rather than a general failure to generalize. The predicted properties reproduce major depositional trends, including Oxfordian lateral continuity and Dogger reservoir compartmentalization, without explicit stratigraphic conditioning. Effective porosity predicted along 1,990 km of interpreted 2D seismic profiles was then integrated into stochastic three-dimensional reservoir modeling, generating 100 equiprobable realizations that represent reservoir heterogeneity and uncertainty while remaining consistent with well observations and regional seismic trends
SN - 0375-6505
ER -