Petrophysically coupled joint inversion of gravity and magnetic data using guided fuzzy clustering and borehole-constrained magnetization directions: Application to the La Primavera geothermal field, Mexico

Gravity and magnetic data are usually inverted separately and integrated only at the interpretation stage, which leaves ambiguous which combination of density and magnetization corresponds to each geological unit. In volcanic geothermal systems two further obstacles arise: lithologies overlap in each property, and remanent magnetization of both polarities invalidates the assumption of induced magnetization. We address them with a two-stage workflow. A magnetic vector inversion guided by borehole paleomagnetic data first fixes the magnetization direction of each cell; a petrophysically coupled joint inversion with guided fuzzy clustering then recovers density, magnetization and their petrophysical units. The coupled problem is solved at three levels that share the objective function and differ only in which unknowns enter the linear system: alternating updates of the cluster centers and memberships, centers solved with the property models, and memberships also solved through Lagrange multipliers. On a synthetic model in which neither property alone separates the units, the two closed-form schemes reduce the error of the anomalous cells by 7–15 percentage points relative to separate inversion and roughly double the agreement of the partition, whereas solving the memberships helps only when the centers are weakly anchored. A Monte Carlo ensemble of 222 inversions shows that the partition is controlled by the prior centers, and a sweep of the assumed magnetization direction shows that it tolerates errors below 20–30°, but not a reversed polarity, which destroys it with a smaller misfit than a 45° error. At La Primavera (Mexico) the joint inversion fits 267 gravity and 530 magnetic data to their noise levels, which the induced direction cannot do. An ablation of the prior information shows that the agreement with 19 of 29 logged well cells comes mainly from the borehole reference model, whereas the coupling sets the unit volumes away from the wells, reducing the cluster nearest to zero contrast from 50 to 29% of the core. The lithic tuff is not resolved: its cluster absorbs the cells that the data leave between the other centers.

Authors

Institutions

Publication Details

Journal
Geothermics
Published
2026-11-17
DOI
https://doi.org/10.1016/j.geothermics.2026.103887
Primary Topic
Geophysical and Geoelectrical Methods
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Petrophysically coupled joint inversion of gravity and magnetic data using guided fuzzy clustering and borehole-constrained magnetization directions: Application to the La Primavera geothermal field, Mexico

F. J. Núñez-Cornũ, Mario Alberto Fuentes-Arreazola, Jonathan Carrillo
Geothermics
Geophysical and Geoelectrical Methods
article

Petrophysically coupled joint inversion of gravity and magnetic data using guided fuzzy clustering and borehole-constrained magnetization directions: Application to the La Primavera geothermal field, Mexico

F. J. Núñez-Cornũ, Mario Alberto Fuentes-Arreazola, Jonathan Carrillo
article en

Abstract

Gravity and magnetic data are usually inverted separately and integrated only at the interpretation stage, which leaves ambiguous which combination of density and magnetization corresponds to each geological unit. In volcanic geothermal systems two further obstacles arise: lithologies overlap in each property, and remanent magnetization of both polarities invalidates the assumption of induced magnetization. We address them with a two-stage workflow. A magnetic vector inversion guided by borehole paleomagnetic data first fixes the magnetization direction of each cell; a petrophysically coupled joint inversion with guided fuzzy clustering then recovers density, magnetization and their petrophysical units. The coupled problem is solved at three levels that share the objective function and differ only in which unknowns enter the linear system: alternating updates of the cluster centers and memberships, centers solved with the property models, and memberships also solved through Lagrange multipliers. On a synthetic model in which neither property alone separates the units, the two closed-form schemes reduce the error of the anomalous cells by 7–15 percentage points relative to separate inversion and roughly double the agreement of the partition, whereas solving the memberships helps only when the centers are weakly anchored. A Monte Carlo ensemble of 222 inversions shows that the partition is controlled by the prior centers, and a sweep of the assumed magnetization direction shows that it tolerates errors below 20–30°, but not a reversed polarity, which destroys it with a smaller misfit than a 45° error. At La Primavera (Mexico) the joint inversion fits 267 gravity and 530 magnetic data to their noise levels, which the induced direction cannot do. An ablation of the prior information shows that the agreement with 19 of 29 logged well cells comes mainly from the borehole reference model, whereas the coupling sets the unit volumes away from the wells, reducing the cluster nearest to zero contrast from 50 to 29% of the core. The lithic tuff is not resolved: its cluster absorbs the cells that the data leave between the other centers.

GeothermicsVol. 143
Universidad de Guadalajara (MX), Centro de Investigación Científica y de Educación Superior de Ensenada (MX)
Centro Mexicano de Innovación en Energía Geotérmica, Consejo Nacional de Ciencia y Tecnología
Geophysical and Geoelectrical Methods
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon