Computational Geosciences最新文献

Seismic imaging is a major challenge in geophysics with broad applications. It involves solving wave propagation equations with absorbing boundary conditions (ABC) multiple times. This drives the need for accurate and efficient numerical methods. This study examines a collection of exponential integration methods, known for their good numerical properties on wave representation, to investigate their efficacy in solving the wave equation with ABC. The purpose of this research is to assess the performance of these methods. We compare a recently proposed Exponential Integration based on Faber polynomials with well-established Krylov exponential methods alongside a high-order Runge-Kutta scheme and low-order classical methods. Through our analysis, we found that the exponential integrator based on the Krylov subspace exhibits the best convergence results among the high-order methods. We also discovered that high-order methods can achieve computational efficiency similar to low-order methods while allowing for considerably larger time steps. Most importantly, the possibility of undertaking large time steps could be used for important memory savings in full waveform inversion imaging problems.

This study introduces two novel DEM-based distributed rainfall-runoff models derived from the existing Hidropixel model: Hidropixel_TUH+ and Hidropixel_DLR. These models account for spatial variations in direct runoff generation, translation, and storage within a watershed, considering spatial variability in rainfall and basin characteristics. In Hidropixel_TUH+, a Triangular Unit Hydrograph (TUH) is determined for each Digital Elevation Model (DEM) pixel and lagged to the watershed outlet based on the travel time from the pixel to the outlet. In Hidropixel_DLR, a hydrograph is estimated for each pixel based on the travel time, which takes translation effects into account. To represent the storage effects, this hydrograph is attenuated by a linear reservoir at each pixel. Both approaches were applied to the Upper Medway catchment (250 km²) in southeastern England, using rainfall data from a rain gauge network. The outcomes revealed that the proposed approaches provided a reasonably accurate prediction of the hydrographs and exhibited notably superior performance compared to the original version of Hidropixel, which has limited capabilities in capturing translation effects. Hidropixel_TUH+ and Hidropixel_DLR predicted peak flows with an average absolute error of 11% and 10%, respectively. The Hidropixel_DLR achieved a more accurate time-to-peak estimation, with an average absolute error of 1 h, compared to the 1.5-h error from Hidropixel_TUH+. Additionally, the Hidropixel_DLR predicted the full direct runoff hydrograph more accurately, achieving an average Nash–Sutcliffe coefficient (NSE) of 0.89, while the Hidropixel_TUH+ had an NSE of approximately 0.84.

Although Artificial Intelligence (AI) projects are common and desired by many institutions and research teams, there are still relatively few success stories of AI in practical use for the Earth science community. Many AI practitioners in Earth science are trapped in the prototyping stage and their results have not yet been adopted by users. Many scientists are still hesitating to use AI in their research routine. This paper aims to capture the landscape of AI-powered geospatial data sciences by discussing the current and upcoming needs of the Earth and environmental community, such as what practical AI should look like, how to realize practical AI based on the current technical and data restrictions, and the expected outcome of AI projects and their long-term benefits and problems. This paper also discusses unavoidable changes in the near future concerning AI, such as the fast evolution of AI foundation models and AI laws, and how the Earth and environmental community should adapt to these changes. This paper provides an important reference to the geospatial data science community to adjust their research road maps, find best practices, boost the FAIRness (Findable, Accessible, Interoperable, and Reusable) aspects of AI research, and reasonably allocate human and computational resources to increase the practicality and efficiency of Earth AI research.

Our research is positioned within the framework of subsurface resource utilization for sustainable economies. We concentrate on modeling the underground single-phase fluid flow affected by geological faults using numerical simulations. The study of such flows is characterized by strong uncertainites in the data defing the problem due to the difficulty of taking precise measurements in the subsoil. We aim to demonstrate the feasibility of a reduced order model that is both reliable and computationally efficient, thereby facilitating the incorporation of uncertainties. We account for the uncertainities of the properties of the rock and the geometry of the fault. The latter is achieved by using a radial basis function mesh deformation method. This approach benefits from a mixed-dimensional framework to model the rock matrix and faults as n and ({n-1}) dimensional domains, allowing for non-conforming meshes. Our primary focus is on a reduced-order model capable of reproducing flow variables across the entire domain. We utilize the Deep Learning Reduced Order Model (DL-ROM), a nonintrusive neural network-based technique, and we compare it against the traditional Proper Orthogonal Decomposition (POD) method across various scenarios. The most relevant contributions of this work are: the proof of concept of the use of neural network for reduced order models for subsoil flow, dealing with non-affine problems and mixed dimensional domain. Additionally, we generalize an existing mesh deformation method for discontinuous deformation maps. Our analysis highlights the capability of reduced order model, highlighting DL-ROM’s capacity to expedite complex analyses with promising accuracy and efficiency, making multi-query analyses with various quantities of interest affordable.

One of the key challenges in assessing, managing and mitigating induced-seismicity hazards related to hydraulic fracturing and fluid injection activities is understanding how geological and operational features influence the likelihood and severity of an event. Geological features point to the pre-existing conditions that affect a well’s susceptibility to generating induced seismicity. In contrast, operational features are controllable and can be engineered to mitigate and minimize potential hazards. In recent years, with increased data availability and the rapid development of machine learning techniques, the application of these statistical tools has been proposed to investigate induced seismicity. However, this raises the question of the performance and interpretability of these methods, which requires thorough investigation. This paper presents the results of a detailed study utilizing data for the Montney region of northeastern British Columbia that investigates the robustness of several machine learning algorithms in predicting induced seismicity likelihood and severity and compares the importance of geological and operational features on the triggering and maximum magnitude of these events. The analyses include seismic monitoring, regional geology and well completions data, and the novel use of geophysical well log data to provide a more comprehensive database of geological features.

Full waveform inversion (FWI), a powerful geophysical technique for subsurface imaging through seismic velocity-model construction, relies on numerical optimization, thus requiring the computation of derivatives for an objective function. This paper proposes a discrete development for accurate computation of the gradient and Hessian-vector product, providing second-order optimization benefits like higher convergence rates and improved resolution. The approach is a promising alternative for computing the gradient and Hessian action in time-domain FWI, applicable to various geophysical problems. Computational costs and memory requirements are comparable to the Adjoint-State Method and more avorable than Automatic Differentiation. While efficient automatic differentiation algorithms have transformed gradient computation in applications like FWI, challenges may arise in 3D due to unforeseen memory allocations. Our approach addresses this by exploring the reverse mode differentiation algorithm, mapping temporary memory allocations and computational complexity. By means of introducing auxiliary fields all involved wavefield evolutions can be carried out with the very same evolution scheme, in this way simplifying the implementation and focusing the performance improvement effort in a single routine thus reducing the maintenance cost of these algorithms, especially when using GPU implementations.

Regional groundwater modelling can provide decision-makers and scientists with valuable information required for the sustainable use and protection of groundwater resources in the future. In order to assess and manage the impact of climate change on regional aquifer systems, numerical groundwater models are required which represent the subsurface structures of aquifers and aquitards in 3D at the regional scale and beyond in the most efficient way. A workflow to clearly generate these structural subsurface representations from a variety of data sources is introduced, applying open-source Geographical Information Systems. The resulting structural models can be used with finite element method-based simulation tools, such as the open-source environment OpenGeoSys. The preparation workflow of the structure model is presented for a large river basin in Germany, indicating the applicability of the method even in a challenging hydrogeological region with several stockworks of dipped and fractured sedimentary aquifers, partially showing significantly changing hydraulic conditions due to natural lateral facies changes.

In this article, we present a novel data assimilation strategy in pore-scale imaging and demonstrate that this makes it possible to robustly address reactive inverse problems incorporating Uncertainty Quantification (UQ). Pore-scale modeling of reactive flow offers a valuable opportunity to investigate the evolution of macro-scale properties subject to dynamic processes in the context of Carbon Capture and Storage (CCS). Yet, they suffer from imaging limitations arising from the associated X-ray microtomography (X-ray (mu )CT) process, which induces discrepancies in the properties estimates. Assessment of the kinetic parameters also raises challenges, as reactive coefficients are critical parameters that can cover a wide range of values. We account for these two issues and ensure reliable calibration of pore-scale modeling, based on dynamical (mu )CT images, by integrating uncertainty quantification in the workflow. The present method is based on a multitasking formulation of reactive inverse problems combining data-driven and physics-informed techniques in calcite dissolution. This allows quantifying morphological uncertainties on the porosity field and estimating reactive parameter ranges through prescribed PDE models, with a latent concentration field, and dynamical (mu )CT observations. The data assimilation strategy relies on sequential reinforcement incorporating successively additional PDE constraints and suitable formulation of the heterogeneous diffusion differential operator leading to enhanced computational efficiency. We provide a robust and unbiased uncertainty quantification by straightforward adaptive weighting of Bayesian Physics-Informed Neural Networks (BPINNs), ensuring reliable micro-porosity changes during geochemical transformations. We demonstrate successful Bayesian Inference in 1D+Time calcite dissolution based on synthetic (mu )CT images with meaningful posterior distribution on the reactive parameters and dimensionless numbers. We eventually apply this framework to a more realistic 2D+Time data assimilation problem involving heterogeneous porosity levels and synthetic (mu )CT dynamical observations.

The numerical simulation of multiphase and multicomponent flows in oil reservoirs is a significant challenge, demanding robust and computationally efficient numerical formulations. Particularly, scenarios with high mobility ratios between injected and resident fluids can lead to Grid Orientation Effects (GOE), where numerical solutions strongly depend on the alignment between flow and computational grid and mobility ratio. This phenomenon relates to an anisotropic distribution in truncation error tied to the numerical approximation of the transport term. Although the oil industry commonly uses linear Two Point Flux Approximation (TPFA) for diffusive fluxes and the First Order Upwind (FOU) method for advective fluxes, both lack rotational invariance and TPFA struggles with non-k-orthogonal grids. This paper proposes a comprehensive cell-centered finite-volume formulation to simulate reservoir oil-water displacements, integrating the classical IMPES (Implicit Pressure Explicit Saturation) segregate approach with unstructured, non-k-orthogonal meshes. Diffusive flux discretization employs a Multipoint Flux Approximation with Harmonic Points (MPFA-H), capable of handling heterogeneous and strongly anisotropic media. A modified second-order Flux Corrected Transport (FCT) approach curbs artificial numerical diffusion for transport term discretization. Additionally, we incorporate a Flow-Oriented Scheme (FOS) for computing low-order and high-order approximations of the numerical fluxes to enhance multidimensional approximation and reduce GOE. The proposed strategy is validated through benchmark problems, yielding precise outcomes with reduced numerical diffusion and GOE effects, underscoring its efficiency for complex reservoir flow simulations.

An ensemble of rigorously history matched reservoir models can help better understand the interactions between heterogeneity and fluid flows, improve forecasting reliability, and locate infill-drilling opportunities to support field development plans. However, developing efficient calibration methods for complex, multi-million cell deep-water models remains a challenge. This paper presents a hierarchical global-local assisted-history matching (AHM) approach with new elements, applied to a complex deep-water reservoir. The method consists of two stages: global and local. In the global stage, the reservoir energy is matched using an evolutionary approach to calibrate the model parameters with build-up and average reservoir pressures. In the local stage, the permeability field is calibrated to production data using a novel streamline-based sensitivity-driven AHM method to ascertain the spatial variability and geologic continuity of local updates. The sensitivity for each streamline is weighted by the water fraction and constrained by a time-of-flight cutoff to focus on water intrusion regions within the near wellbore region. The proposed method is field-tested in a complex deep-water reservoir. The evolutionary approach generates an ensemble of models with well-matched oil production rates and build-up/reservoir pressure using global model parameters. Local updates using streamline-based gradients are then conducted to match the water cut for each ensemble member while maintaining overall pressure match quality. Results show that the hierarchical AHM method significantly reduces the data misfit and is well-suited to primary recovery in a deep-water setting with few producers and under the influence of mild/weak aquifers.