Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Probabilistic Joint Recovery Method for CO$_2$ Plume Monitoring

Zijun Deng, Rafael Orozco, Abhinav Prakash Gahlot, Felix J. Herrmann

TL;DR

The paper tackles uncertainty in time-lapse seismic monitoring of $CO_2$ plumes for CCS. It introduces a probabilistic joint recovery framework (pJRM) that uses a Shared Generative Model $G_\theta$ with per-survey encodings $q_{\phi_i}$ to infer posterior distributions from forward operators $\mathbf{A}_i$ and measurements $\mathbf{y}_i$. Compared with the probabilistic independent baseline (pIRM), pJRM exploits shared structure across surveys and employs a weak formulation to amortize costly forward evaluations, yielding time-lapse images from differences of posterior means $\mathbb{E}[\mathbf{x}_i]$. In synthetic CCS scenarios with varying survey counts, pJRM improves plume reconstruction and reduces uncertainty as more surveys are included, demonstrating practical value for risk-aware CCS monitoring.

Abstract

Reducing CO$_2$ emissions is crucial to mitigating climate change. Carbon Capture and Storage (CCS) is one of the few technologies capable of achieving net-negative CO$_2$ emissions. However, predicting fluid flow patterns in CCS remains challenging due to uncertainties in CO$_2$ plume dynamics and reservoir properties. Building on existing seismic imaging methods like the Joint Recovery Method (JRM), which lacks uncertainty quantification, we propose the Probabilistic Joint Recovery Method (pJRM). By estimating posterior distributions across surveys using a shared generative model, pJRM provides uncertainty information to improve risk assessment in CCS projects.

Probabilistic Joint Recovery Method for CO$_2$ Plume Monitoring

TL;DR

The paper tackles uncertainty in time-lapse seismic monitoring of plumes for CCS. It introduces a probabilistic joint recovery framework (pJRM) that uses a Shared Generative Model with per-survey encodings to infer posterior distributions from forward operators and measurements . Compared with the probabilistic independent baseline (pIRM), pJRM exploits shared structure across surveys and employs a weak formulation to amortize costly forward evaluations, yielding time-lapse images from differences of posterior means . In synthetic CCS scenarios with varying survey counts, pJRM improves plume reconstruction and reduces uncertainty as more surveys are included, demonstrating practical value for risk-aware CCS monitoring.

Abstract

Reducing CO emissions is crucial to mitigating climate change. Carbon Capture and Storage (CCS) is one of the few technologies capable of achieving net-negative CO emissions. However, predicting fluid flow patterns in CCS remains challenging due to uncertainties in CO plume dynamics and reservoir properties. Building on existing seismic imaging methods like the Joint Recovery Method (JRM), which lacks uncertainty quantification, we propose the Probabilistic Joint Recovery Method (pJRM). By estimating posterior distributions across surveys using a shared generative model, pJRM provides uncertainty information to improve risk assessment in CCS projects.

Paper Structure

This paper contains 9 sections, 3 equations, 3 figures, 1 algorithm.

Table of Contents

  1. Introduction
  2. Previous Work
  3. Method
  4. Probabilistic Joint Recovery Model
  5. Probabilistic Independent Recovery Model
  6. Weak formulation
  7. Synthetic case study
  8. Conclusions
  9. Acknowledgments

Figures (3)

  • Figure 1: Synthetic case study. (a) Earth model with CCS injection site and unknown CO$_2$ plume. (b) Simulated post-stack data at first survey. (c) Simulated post-stack data at second survey.
  • Figure 2: Comparison of time-lapse Images. (a) Time-lapse image with Independent pIRM. (b) Ground truth time-lapse difference. (c) Time-lapse image with pJRM and 2 surveys. (d) Time-lapse image with pJRM and 6 surveys.
  • Figure 3: Comparison of uncertainty and errors. (a) Uncertainty of time-lapse image w/ independent method pIRM and 2 surveys. (b) Error of time-lapse image w/ pIRM and 2 surveys. (c) Uncertainty of time-lapse image w/ joint method pJRM and 2 surveys. (d) Error of time-lapse image w/ pJRM and 2 surveys. (e) Uncertainty of time-lapse image w/ joint method pJRM and 6 surveys. (f) Error of time-lapse image w/ pJRM and 6 surveys.