FRANCISCOESCAMILLA

I am FRANCISCO ESCAMILLA, a geophysical data scientist and computational seismologist specializing in advancing seismic imaging and hazard prediction through innovative data augmentation frameworks for wave scattering phenomena. With a Ph.D. in Wave Physics and Inverse Theory (Stanford University, 2020) and a Senior Research Fellowship at the Swiss Seismological Service (ETH Zurich, 2021–2024), I have pioneered hybrid methodologies that bridge physics-driven simulations and machine learning to overcome data scarcity in complex geological settings. As the Chief Architect of the Scattering Augmentation Lab (SAL) and Principal Investigator of the EU-funded QUAKE-CORE project, I develop scalable solutions to enhance seismic dataset diversity, resolution, and interpretability. My work on stochastic scattering augmentation received the 2023 American Geophysical Union’s Keiiti Aki Award and supports UNESCO’s global earthquake early-warning initiatives.

Research Motivation

Seismic wave scattering—heterogeneous interactions of seismic waves with subsurface structures—is critical for hydrocarbon exploration, volcanic monitoring, and megathrust earthquake studies. However, field data acquisition faces three fundamental limitations:

  1. Data Sparsity: High-fidelity scattering labels are scarce due to prohibitive costs of deep-borehole deployments and controlled-source experiments.

  2. Noise-Dominance: Ambient noise contaminates 60–80% of low-magnitude event recordings, masking subtle scattering signatures.

  3. Computational Intractability: Full-waveform inversion (FWI) of scattering requires petascale resources, limiting real-time applications.

My mission is to create physics-constrained data augmentation pipelines that synthesize realistic scattering waveforms, enabling robust AI-driven subsurface characterization.

Methodological Framework

My research integrates generative adversarial networks (GANs), wave-equation-constrained optimization, and quantum-enhanced sampling:

1. Stochastic Scattering GANs (SS-GAN)

  • Developed GeoAugment, a generative framework that:

    • Synthesizes scattering waveforms by training on 3D elastic wave equations and stochastic media properties (e.g., correlation lengths, velocity fluctuations).

    • Achieves 92% structural similarity (SSIM) to field data from the San Andreas Fault Observatory at Depth (SAFOD).

    • Reduces FWI misfit by 40% in oil reservoir models (validated with Chevron’s Permian Basin datasets).

  • Licensed to Schlumberger for augmenting training data for their DELFI cognitive E&P platform.

2. Physics-Informed Neural Operators (PINO-Scatter)

  • Designed PINO-Scatter, a hybrid neural operator:

    • Embeds viscoelastic wave equations into transformer architectures to extrapolate sparse field data into full scattering wavefields.

    • Predicts S-wave coda envelopes with 0.85 Pearson correlation across 10–50 Hz bandwidths (Geophysical Journal International, 2024).

    • Enabled 30x faster probabilistic seismic hazard analysis (PSHA) for Tokyo’s Metropolitan Earthquake Preparedness Project.

3. Quantum Monte Carlo Scattering (QMC-Scatter)

  • Pioneered QMC-Scatter, a quantum-classical hybrid workflow:

    • Encodes subsurface media uncertainty into 16-qubit quantum states for rapid scattering path sampling.

    • Solves 2D acoustic-elastic coupling problems on IBM’s Heron processors with 70% faster convergence than classical Monte Carlo.

    • Maps scattering kernels for Mars’ Cerberus Fossae region using NASA’s InSight mission data.

Ethical and Technical Innovations

  1. Open Seismic Science

    • Launched ScatterBase, an open repository of 100,000+ augmented scattering waveforms with PyTorch/TensorFlow loaders.

    • Authored the FAIR Scattering Standards to ensure reproducibility in AI-geophysics research.

  2. Disaster Resilience

    • Developed ScatterAlert, a real-time augmentation system that enhances low-quality seismic signals for early tsunami warnings (deployed in Indonesia and Chile).

    • Collaborated with the IAEA to detect clandestine nuclear tests via augmented scattering coda waves.

  3. Sustainable Computation

    • Introduced GreenFWI, a data-augmented FWI framework reducing carbon footprint by 65% through optimized gradient calculations.

    • Advocated for Ethical Augmentation Guidelines to prevent misuse in geopolitical resource disputes.

Global Impact and Future Visions

  • 2023–2025 Milestones:

    • Accelerated geothermal site assessments by 50% in Iceland’s Krafla Volcano region using augmented scattering templates.

    • Trained 800+ geoscientists through the Scattering Augmentation Mastery Program (SAMP).

    • Partnered with Google to integrate QMC-Scatter into their Quantum Cloud for global seismic risk dashboards.

  • Vision 2026–2030:

    • Exascale Augmentation: Real-time scattering synthesis for planetary-scale FWI on next-generation supercomputers.

    • Bio-Inspired Scattering: Mimicking bat echolocation strategies to design self-optimizing augmentation networks.

    • Interstellar Seismology: Adapting augmentation pipelines to analyze ice-quake scattering on Europa and Enceladus.

By treating seismic scattering not as noise but as an information goldmine, I aim to transform geophysical data scarcity into abundance—empowering humanity to forecast disasters, harness Earth’s resources, and explore extraterrestrial frontiers with unprecedented clarity.

Seismic Data Solutions

Expertise in data curation, modeling, and validation for enhanced seismic analysis and interpretation.

Model Development

Utilizing physics-GAN for generator constraints and energy conservation checks in seismic modeling processes.

Abstract image featuring wavy, layered patterns resembling geological formations or sedimentary rock layers with a smooth, flowing texture and intermixed dark and light brown tones.
Abstract image featuring wavy, layered patterns resembling geological formations or sedimentary rock layers with a smooth, flowing texture and intermixed dark and light brown tones.
Validation Services

Comparative analysis of imaging accuracy and microseismic detection sensitivity in observational datasets.

Fine-tuning encodes seismology tokens for integration with advanced interpretation suites and enhanced insights.

API Integration
The image features a detailed pattern of natural rock formations with parallel lines and layered textures. Shades of brown, beige, and black give the surface a rugged, weathered look, with visible variations in color due to sedimentation.
The image features a detailed pattern of natural rock formations with parallel lines and layered textures. Shades of brown, beige, and black give the surface a rugged, weathered look, with visible variations in color due to sedimentation.
Layers of sedimentary rock formation with visible strata patterns. The texture appears rugged and weathered, showing shades of gray, brown, and hints of rust. The geological structure suggests erosion and time-worn surfaces.
Layers of sedimentary rock formation with visible strata patterns. The texture appears rugged and weathered, showing shades of gray, brown, and hints of rust. The geological structure suggests erosion and time-worn surfaces.

Seismic Integration

Merging seismic catalogs with oil well logs for analysis.

Aerial view of a textured sandy surface with wave-like patterns and undulations. The sand appears to be slightly damp, creating a mixture of light and dark areas due to shadows and variations in the surface.
Aerial view of a textured sandy surface with wave-like patterns and undulations. The sand appears to be slightly damp, creating a mixture of light and dark areas due to shadows and variations in the surface.
Model Development

Physics-GAN with wave equation constraints and energy checks.

A seascape featuring two offshore oil rigs positioned in the calm sea. The larger rig is on the right, and a smaller one is on the left, both set against a clear blue sky. In the foreground, a stone jetty separates the water from the viewer.
A seascape featuring two offshore oil rigs positioned in the calm sea. The larger rig is on the right, and a smaller one is on the left, both set against a clear blue sky. In the foreground, a stone jetty separates the water from the viewer.
Validation Process

Comparing imaging accuracy and testing microseismic detection sensitivity.

Two distinct sections of ocean waves from a top-down perspective, with the left side showing deeper blue, turbulent water and the right side displaying lighter, foamy waves over shallower green water.
Two distinct sections of ocean waves from a top-down perspective, with the left side showing deeper blue, turbulent water and the right side displaying lighter, foamy waves over shallower green water.
The image depicts a close-up view of cracked red earth or a similar surface, featuring deep and irregular black fissures and bright white highlights emphasizing the texture.
The image depicts a close-up view of cracked red earth or a similar surface, featuring deep and irregular black fissures and bright white highlights emphasizing the texture.
API Role

Fine-tuning seismology tokens and integrating with interpretation suites.

Data Curation

Labeling attenuation and framing wavefield solutions as inputs.

Innovative Data Solutions

Specializing in seismic data curation and advanced modeling techniques for enhanced geophysical insights.

An aerial view of a mining site with layered earth tones and a long, dark road or conveyor belt running horizontally across the center. Earth-moving equipment is visible near the bottom right, amidst large, sweeping patterns in the terrain that suggest excavation or quarrying activity.
An aerial view of a mining site with layered earth tones and a long, dark road or conveyor belt running horizontally across the center. Earth-moving equipment is visible near the bottom right, amidst large, sweeping patterns in the terrain that suggest excavation or quarrying activity.
A small 3D model represents a section of terrain with green grass and a body of water. On the grassy area, there is a round, white robotic device with red and black accents. The surface of the water reflects the light, creating a realistic appearance.
A small 3D model represents a section of terrain with green grass and a body of water. On the grassy area, there is a round, white robotic device with red and black accents. The surface of the water reflects the light, creating a realistic appearance.
Insights

Advanced data curation and modeling for seismology.

© 2025. All rights reserved.