Hossein Ghadjari
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Hossein Ghadjari has a total of 67 co-authors. Here are the top co-authors by number of shared publications:
Susan Skone
Department of Geomatics Engineering University of Calgary Calgary Alberta Canada
5 shared publications- David J. Knudsen
Department of Physics and Astronomy University of Calgary Calgary AB Canada
3 shared publications - Hersh Gilbert
University of Calgary
3 shared publications - Kamyar Azizzadenesheli
NVIDIA Corporation
3 shared publications Pejman Shahsavari
University of Calgary
3 shared publications- Alfred Mallet
Space Sciences Laboratory, University of California, Berkeley, CA 94720, USA
2 shared publications
Latest Publications
- Open Access
- 14 January 2026
- Published by Oxford University Press (OUP) in Geophysical Journal International
- Vol. 244 (3)
A Fourier neural operator surrogate model for nonlinear electrical resistivity tomography
SUMMARY: Electrical resistivity tomography (ERT) is used to infer the subsurface resistivity structure. ERT requires solving a nonlinear inverse problem that is often approximated as linear to reduce computational time. However, the approximation requires assumptions that cause limitations for the data analysis. Most of the computational time is due to the forward problem that requires solving the Poisson equation. Recently, similar forward problems have been shown to be replaceable with a surrogate model of lower computational cost. We present a geoelectric surrogate based on Fourier neural operators and demonstrate a successful application in nonlinear inversion. The standard deviation of the surrogate prediction error for unseen samples is $<$5 per cent. Furthermore, the surrogate reduces computational time by over three orders of magnitude per realization, enabling ERT for previously intractable settings. We apply the surrogate in Markov chain Monte Carlo inversion of simulated data. The results resolve sharp resistivity changes with plausible uncertainties.
- 14 January 2026
- Published by Oxford University Press (OUP) in Geophysical Journal International
- Vol. 244 (3)
Open AccessA Fourier neural operator surrogate model for nonlinear electrical resistivity tomography
SUMMARY: Electrical resistivity tomography (ERT) is used to infer the subsurface resistivity structure. ERT requires solving a nonlinear inverse problem that is often approximated as linear to reduce computational time. However, the approximation requires assumptions that cause limitations for the data analysis. Most of the computational time is due to the forward problem that requires solving the Poisson equation. Recently, similar forward problems have been shown to be replaceable with a surrogate model of lower computational cost. We present a geoelectric surrogate based on Fourier neural operators and demonstrate a successful application in nonlinear inversion. The standard deviation of the surrogate prediction error for unseen samples is $<$5 per cent. Furthermore, the surrogate reduces computational time by over three orders of magnitude per realization, enabling ERT for previously intractable settings. We apply the surrogate in Markov chain Monte Carlo inversion of simulated data. The results resolve sharp resistivity changes with plausible uncertainties.
- Open Access
Hilbert Neural Operator: Operator Learning in the Analytic Signal Domain
Neural operators have emerged as a powerful, data-driven paradigm for learning solution operators of partial differential equations (PDEs). State-of-the-art architectures, such as the Fourier Neural Operator (FNO), have achieved remarkable success by performing convolutions in the frequency domain, making them highly effective for a wide range of problems. However, this method has some limitations, including the periodicity assumption of the Fourier transform. In addition, there are other methods of analysing a signal, beyond phase and amplitude perspective, and provide us with other useful information to learn an effective network. We introduce the \textbf{Hilbert Neural Operator (HNO)}, a new neural operator architecture to address some advantages by incorporating a strong inductive bias from signal processing. HNO operates by first mapping the input signal to its analytic representation via the Hilbert transform, thereby making instantaneous amplitude and phase information explicit features for the learning process. The core learnable operation -- a spectral convolution -- is then applied to this Hilbert-transformed representation. We hypothesize that this architecture enables HNO to model operators more effectively for causal, phase-sensitive, and non-stationary systems. We formalize the HNO architecture and provide the theoretical motivation for its design, rooted in analytic signal theory.
Open AccessHilbert Neural Operator: Operator Learning in the Analytic Signal Domain
Neural operators have emerged as a powerful, data-driven paradigm for learning solution operators of partial differential equations (PDEs). State-of-the-art architectures, such as the Fourier Neural Operator (FNO), have achieved remarkable success by performing convolutions in the frequency domain, making them highly effective for a wide range of problems. However, this method has some limitations, including the periodicity assumption of the Fourier transform. In addition, there are other methods of analysing a signal, beyond phase and amplitude perspective, and provide us with other useful information to learn an effective network. We introduce the \textbf{Hilbert Neural Operator (HNO)}, a new neural operator architecture to address some advantages by incorporating a strong inductive bias from signal processing. HNO operates by first mapping the input signal to its analytic representation via the Hilbert transform, thereby making instantaneous amplitude and phase information explicit features for the learning process. The core learnable operation -- a spectral convolution -- is then applied to this Hilbert-transformed representation. We hypothesize that this architecture enables HNO to model operators more effectively for causal, phase-sensitive, and non-stationary systems. We formalize the HNO architecture and provide the theoretical motivation for its design, rooted in analytic signal theory.
- Open Access
- 18 July 2025
- Published by EDP Sciences in Journal of Space Weather and Space Climate
- Vol. 15, 30
Intermittency in the integrated power of ionospheric density fluctuations
The occurrence of ionospheric irregularities poses a significant challenge to Global Navigation Satellite Systems (GNSS) by disrupting signal propagation and causing loss of lock (LOL) events. This study investigates the integrated power spectral density (PSD) of electron density fluctuations at various scales using nearly ten years of Swarm satellite data. The analysis focuses on the post-sunset equatorial ionosphere, a region prone to ionospheric irregularities. GPS receivers on Swarm A and C, orbiting at altitudes between 430 and 460 km, suffer more frequent loss of navigational capability (LNC) than Swarm B (at 530 km). The observed power-law distribution of integrated power at small spatial scales (<30 km) points to scale-free behavior, which is a hallmark of complex systems and may be associated with self-organized criticality (SOC). The study also employs multifractal detrended fluctuation analysis (MFDFA) to demonstrate the multifractal and intermittent nature of the fluctuations. The findings highlight the importance of intermittency and strong bursts in understanding the LNC and LOL events in the ionosphere. This research contributes to a deeper understanding of the dynamics of ionospheric irregularities and offers potential for improved forecasting and mitigation of their impact on GNSS.
- 18 July 2025
- Published by EDP Sciences in Journal of Space Weather and Space Climate
- Vol. 15, 30
Open AccessIntermittency in the integrated power of ionospheric density fluctuations
The occurrence of ionospheric irregularities poses a significant challenge to Global Navigation Satellite Systems (GNSS) by disrupting signal propagation and causing loss of lock (LOL) events. This study investigates the integrated power spectral density (PSD) of electron density fluctuations at various scales using nearly ten years of Swarm satellite data. The analysis focuses on the post-sunset equatorial ionosphere, a region prone to ionospheric irregularities. GPS receivers on Swarm A and C, orbiting at altitudes between 430 and 460 km, suffer more frequent loss of navigational capability (LNC) than Swarm B (at 530 km). The observed power-law distribution of integrated power at small spatial scales (<30 km) points to scale-free behavior, which is a hallmark of complex systems and may be associated with self-organized criticality (SOC). The study also employs multifractal detrended fluctuation analysis (MFDFA) to demonstrate the multifractal and intermittent nature of the fluctuations. The findings highlight the importance of intermittency and strong bursts in understanding the LNC and LOL events in the ionosphere. This research contributes to a deeper understanding of the dynamics of ionospheric irregularities and offers potential for improved forecasting and mitigation of their impact on GNSS.
- Open Access
- 2 July 2025
- Published by Wiley
A Fourier Neural Operator Surrogate Model for nonlinear Electrical Resistivity Tomography
Electrical resistivity tomography (ERT) is used to infer the subsurface resistivity structure. ERT requires solving a nonlinear inverse problem that is often approximated as linear to reduce computational time. However, the approximation requires assumptions that cause limitations for the data analysis. Most of the computational time is due to the forward problem that requires solving the Poisson equation. Recently, similar forward problems have been shown to be replaceable with a surrogate model of lower computational cost. We present a geoelectric surrogate based on Fourier Neural Operators (FNO) and demonstrate a successful application in nonlinear inversion. The standard deviation of Surrogate prediction errors for unseen samples are $<$5\%. Furthermore, the surrogate reduces computational time by over three orders of magnitude per realization, enabling ERT for previously intractable settings. We apply the surrogate in Markov chain Monte Carlo (MCMC) inversion of simulated data. The results resolve sharp resistivity changes with plausible uncertainties.
- 2 July 2025
- Published by Wiley
Open AccessA Fourier Neural Operator Surrogate Model for nonlinear Electrical Resistivity Tomography
Electrical resistivity tomography (ERT) is used to infer the subsurface resistivity structure. ERT requires solving a nonlinear inverse problem that is often approximated as linear to reduce computational time. However, the approximation requires assumptions that cause limitations for the data analysis. Most of the computational time is due to the forward problem that requires solving the Poisson equation. Recently, similar forward problems have been shown to be replaceable with a surrogate model of lower computational cost. We present a geoelectric surrogate based on Fourier Neural Operators (FNO) and demonstrate a successful application in nonlinear inversion. The standard deviation of Surrogate prediction errors for unseen samples are $<$5\%. Furthermore, the surrogate reduces computational time by over three orders of magnitude per realization, enabling ERT for previously intractable settings. We apply the surrogate in Markov chain Monte Carlo (MCMC) inversion of simulated data. The results resolve sharp resistivity changes with plausible uncertainties.
- Open Access
- 2 May 2025
- Published by Wiley
A Fourier Neural Operator Surrogate Model for nonlinear Electrical Resistivity Tomograph
Original Article Electrical resistivity tomography (ERT) is used to infer the subsurface resistivity structure. ERT requires solving a nonlinear inverse problem that is often approximated as linear to reduce computational time. However, the approximation requires assumptions that cause limitations for the data analysis. Most of the computational time is due to the forward problem that requires solving the Poisson equation. Recently, similar forward problems have been shown to be replaceable with a surrogate model of lower computational cost. We present a geoelectric surrogate based on Fourier Neural Operators (FNO) and demonstrate a successful application in nonlinear inversion. The standard deviation of Surrogate prediction errors for unseen samples are <5%. Furthermore, the surrogate reduces computational time by over two orders of magnitude, enabling ERT for previously intractable settings. We apply the surrogate in Markov chain Monte Carlo (MCMC) inversion of simulated data. The results resolve sharp resistivity changes with plausible uncertainties.
- 2 May 2025
- Published by Wiley
Open AccessA Fourier Neural Operator Surrogate Model for nonlinear Electrical Resistivity Tomograph
Original Article Electrical resistivity tomography (ERT) is used to infer the subsurface resistivity structure. ERT requires solving a nonlinear inverse problem that is often approximated as linear to reduce computational time. However, the approximation requires assumptions that cause limitations for the data analysis. Most of the computational time is due to the forward problem that requires solving the Poisson equation. Recently, similar forward problems have been shown to be replaceable with a surrogate model of lower computational cost. We present a geoelectric surrogate based on Fourier Neural Operators (FNO) and demonstrate a successful application in nonlinear inversion. The standard deviation of Surrogate prediction errors for unseen samples are <5%. Furthermore, the surrogate reduces computational time by over two orders of magnitude, enabling ERT for previously intractable settings. We apply the surrogate in Markov chain Monte Carlo (MCMC) inversion of simulated data. The results resolve sharp resistivity changes with plausible uncertainties.
- Open Access
- 18 March 2025
- Published by Copernicus GmbH
Nonlinear Electrical Resistivity Tomography with a Fourier Neural Operator Surrogate Model
Electrical resistivity tomography (ERT) is commonly applied for shallow subsurface imaging. Inversion techniques generate images of the subsurface resistivity structure to interpret the data, with applications including the imaging of permafrost soils. While linearized inversion is a common method, nonlinear treatment provides advantages in terms of parametrization and model selection. However, it often incurs prohibitive computational costs. Markov Chain Monte Carlo (MCMC) methods offer nonlinear uncertainty quantification for ERT, where the computational cost is dominated by the forward model evaluations. Surrogate models advance the physics forward model with a considerable speedup; therefore, they have the potential to enable MCMC applications for inverse problems that were not previously possible. We introduce a surrogate forward model for 2D ERT based on a Fourier Neural Operator (FNO). This model leverages the FNO's capability to learn and generalize mappings between infinite-dimensional function spaces, making it particularly suitable for solving PDE-driven problems like ERT. Based on the inputs of electrode geometry and subsurface resistivity distribution, FNO predicts potentials from which apparent resistivities are computed. This process reduces evaluation times of a subsurface resistivity distribution from seconds to milliseconds with prediction errors below 5%. This efficiency gain enables applying the FNO in MCMC sampling. We show several examples of MCMC sampling results with simulated data for pole-dipole arrays and realistic subsurface models. The subsurface parametrization of resistivity considers irregular grids based on Gaussian random fields or Voronoi cells. The results demonstrate that nonlinear inversion and uncertainty quantification are computationally feasible for typical field survey scales.
- 18 March 2025
- Published by Copernicus GmbH
Open AccessNonlinear Electrical Resistivity Tomography with a Fourier Neural Operator Surrogate Model
Electrical resistivity tomography (ERT) is commonly applied for shallow subsurface imaging. Inversion techniques generate images of the subsurface resistivity structure to interpret the data, with applications including the imaging of permafrost soils. While linearized inversion is a common method, nonlinear treatment provides advantages in terms of parametrization and model selection. However, it often incurs prohibitive computational costs. Markov Chain Monte Carlo (MCMC) methods offer nonlinear uncertainty quantification for ERT, where the computational cost is dominated by the forward model evaluations. Surrogate models advance the physics forward model with a considerable speedup; therefore, they have the potential to enable MCMC applications for inverse problems that were not previously possible. We introduce a surrogate forward model for 2D ERT based on a Fourier Neural Operator (FNO). This model leverages the FNO's capability to learn and generalize mappings between infinite-dimensional function spaces, making it particularly suitable for solving PDE-driven problems like ERT. Based on the inputs of electrode geometry and subsurface resistivity distribution, FNO predicts potentials from which apparent resistivities are computed. This process reduces evaluation times of a subsurface resistivity distribution from seconds to milliseconds with prediction errors below 5%. This efficiency gain enables applying the FNO in MCMC sampling. We show several examples of MCMC sampling results with simulated data for pole-dipole arrays and realistic subsurface models. The subsurface parametrization of resistivity considers irregular grids based on Gaussian random fields or Voronoi cells. The results demonstrate that nonlinear inversion and uncertainty quantification are computationally feasible for typical field survey scales.
- Open Access
- 31 July 2023
- Published by American Astronomical Society
Next Generation Machine to Study Heliophysics in the Laboratory
Whitepaper #099 in the Decadal Survey for Solar and Space Physics (Heliophysics) 2024-2033. Main topics: basic research; infrastructure/workforce/other programmatic. Additional topics: ground-based missions/projects; laboratory space plasma physics; research tools and […]
- 31 July 2023
- Published by American Astronomical Society
Open AccessNext Generation Machine to Study Heliophysics in the Laboratory
Whitepaper #099 in the Decadal Survey for Solar and Space Physics (Heliophysics) 2024-2033. Main topics: basic research; infrastructure/workforce/other programmatic. Additional topics: ground-based missions/projects; laboratory space plasma physics; research tools and […]
- Open Access
- 31 July 2023
- Published by American Astronomical Society
Enabling Discoveries in Heliospheric Science through Laboratory Plasma Experiments
Whitepaper #236 in the Decadal Survey for Solar and Space Physics (Heliophysics) 2024-2033. Main topics: basic research; infrastructure/workforce/other programmatic. Additional topics: fundamental space plasma theory; state of the profession; laboratory space plasma […]
- 31 July 2023
- Published by American Astronomical Society
Open AccessEnabling Discoveries in Heliospheric Science through Laboratory Plasma Experiments
Whitepaper #236 in the Decadal Survey for Solar and Space Physics (Heliophysics) 2024-2033. Main topics: basic research; infrastructure/workforce/other programmatic. Additional topics: fundamental space plasma theory; state of the profession; laboratory space plasma […]
- Open Access
- 13 July 2023
- Published by American Geophysical Union (AGU) in Geophysical Research Letters
- Vol. 50 (14)
Post‐Sunset Field‐Line Resonances at Equatorial Latitudes Observed by Swarm
In this study employing data from the Swarm satellites at an altitude of ≈500 km, we present evidence of field‐line resonances (FLRs) in the Pc2‐3 (0.02–0.2 Hz) frequency band in the undisturbed post‐sunset region of the equatorial ionosphere. We identify 26 events extending to magnetic latitudes as low as 2°, which was previously thought to be too low for FLRs to exist due to ionospheric damping. Identification of FLRs is supported by a narrow‐band fundamental and first harmonics, an oscillating field‐aligned Poynting vector, a relative phase shift between E and δB of ≈90°, and in one case observation of one wave in conjugate hemispheres by two satellites separated by 4 min. We show that, contrary to the non‐uniform trend of ground‐based observations, the fundamental frequencies of these occurrences decline monotonically with increasing field‐line length. Their frequencies differ from those of ground‐based magnetometer observations reported in the literature.
- 13 July 2023
- Published by American Geophysical Union (AGU) in Geophysical Research Letters
- Vol. 50 (14)
Open AccessPost‐Sunset Field‐Line Resonances at Equatorial Latitudes Observed by Swarm
In this study employing data from the Swarm satellites at an altitude of ≈500 km, we present evidence of field‐line resonances (FLRs) in the Pc2‐3 (0.02–0.2 Hz) frequency band in the undisturbed post‐sunset region of the equatorial ionosphere. We identify 26 events extending to magnetic latitudes as low as 2°, which was previously thought to be too low for FLRs to exist due to ionospheric damping. Identification of FLRs is supported by a narrow‐band fundamental and first harmonics, an oscillating field‐aligned Poynting vector, a relative phase shift between E and δB of ≈90°, and in one case observation of one wave in conjugate hemispheres by two satellites separated by 4 min. We show that, contrary to the non‐uniform trend of ground‐based observations, the fundamental frequencies of these occurrences decline monotonically with increasing field‐line length. Their frequencies differ from those of ground‐based magnetometer observations reported in the literature.
- Open Access
- 15 May 2023
- Published by Copernicus GmbH
Probability distribution of integrated power of equatorial ionosphere plasma density fluctuations measured by the Swarm Langmuir probes
Ionospheric irregularities are structures or fluctuations of plasma density having different scale sizes. These irregularities can disrupt radio waves and produce errors in space-based or ground-based technologies, which depend on the GNSS/ GPS signals. Post-sunset ionospheric plasma irregularities are a common characteristic of the equatorial ionosphere. These irregularities, associated with plasma bubbles, are defined as strong density depletions relative to the background plasma as determined by in situ measurements. Finding a system parameter's probability distribution function (PDF) can lead us to understand the system's underlying physics. In this study, we investigate the probability distribution of the integrated power of post-sunset plasma density irregularities in the equatorial ionosphere measured with the Langmuir probes on Swarm C in four different frequency bands between 0-1 Hz for the entire Swarm mission. We find evidence of "heavy tail" distribution in the PDFs, indicating the system's complexity and self-organized criticality. Moreover, we study the relation between Integrated power and different geomagnetic indices, e.g. F10.7 and sunspot number, to find the potential drivers of severe events. While we find no obvious driver of individual events, we find a strong solar cycle dependence in their occurrence.
- 15 May 2023
- Published by Copernicus GmbH
Open AccessProbability distribution of integrated power of equatorial ionosphere plasma density fluctuations measured by the Swarm Langmuir probes
Ionospheric irregularities are structures or fluctuations of plasma density having different scale sizes. These irregularities can disrupt radio waves and produce errors in space-based or ground-based technologies, which depend on the GNSS/ GPS signals. Post-sunset ionospheric plasma irregularities are a common characteristic of the equatorial ionosphere. These irregularities, associated with plasma bubbles, are defined as strong density depletions relative to the background plasma as determined by in situ measurements. Finding a system parameter's probability distribution function (PDF) can lead us to understand the system's underlying physics. In this study, we investigate the probability distribution of the integrated power of post-sunset plasma density irregularities in the equatorial ionosphere measured with the Langmuir probes on Swarm C in four different frequency bands between 0-1 Hz for the entire Swarm mission. We find evidence of "heavy tail" distribution in the PDFs, indicating the system's complexity and self-organized criticality. Moreover, we study the relation between Integrated power and different geomagnetic indices, e.g. F10.7 and sunspot number, to find the potential drivers of severe events. While we find no obvious driver of individual events, we find a strong solar cycle dependence in their occurrence.