GeoFWI3D

Large-scale 3D velocity models for deep-learning full waveform inversion (FWI) and other seismic processing workflows.

Overview

GeoFWI3D provides many synthetic 3D Earth models, each stored as raw float32 (little-endian) binaries. Per model you get:

File	Contents
vp3d.bin	P-wave velocity
image3d.bin	Seismic image
rgt3d.bin	Relative geologic time (RGT)
fault3d.bin	Fault index mask

Each volume has shape 96 × 96 × 96 in (X, Y, Z) order. On disk the layout is contiguous C-order over those dimensions.

Download

Compressed archives are hosted on Box:

Download GeoFWI3D data

Each models_batch_*.tar.gz contains 1000 models. Place the archive files in the same directory as this repository, then extract.

Extract

From the repository root:

./extract_models.sh

This creates allmodels/ and extracts every models_batch_*.tar.gz into it.

Directory layout

After extraction, models look like:

allmodels/
├── model_0000/
│   ├── image3d.bin
│   ├── vp3d.bin
│   ├── rgt3d.bin
│   └── fault3d.bin
├── model_0001/
│   └── ...
└── ...

Folder names use four-digit zero padding: model_0000, model_0001, …

Quick start

1. Install Python dependencies used by the example notebook: NumPy, Matplotlib, and scikit-image (for marching_cubes).
2. Run Jupyter with working directory quick_start/ so from plotting import plot3d works. Open read_data.ipynb: it defines read_models, sets data_root / model_folders / shape, and walks through loading and plotting.

image1, vp1, rgt1, fault1 = read_models(data_root, model_folders, 1872)

fig, ax = plt.subplots(
    2, 2,
    figsize=(10, 10),
    subplot_kw={"projection": "3d"},
    constrained_layout=True,
)
axes = ax.ravel()

data_list = [image1, vp1, rgt1, fault1]
titles = ["Seismic image", "Velocity model", "Relative geologic time", "Fault model"]
cmaps = ["gray", "jet", "viridis", "jet"]

for ax, data, title, cmap_name in zip(axes, data_list, titles, cmaps):
    ax.set_title(title, fontsize=16, fontweight="bold", pad=15)
    plt.sca(ax)
    plot3d(
        data.T,
        cmap=cmap_name,
        frames=[45, 73, 60],
        ifnewfig=False,
        showf=False,
        close=False,
        ifinside=False,
    )

plt.show()

fault3d.bin stores a fault index per voxel. To plot a single fault (here index 3):

fault_mask = fault1.T == 3
plot3d(
    fault_mask.astype(np.float32),
    cmap="Reds",
    frames=[25, 73, 12],
    ifnewfig=True,
    showf=False,
    close=False,
    ifinside=False,
)

Salt bodies have RGT = 0 in rgt3d.bin. Mask and plot with:

salt_mask = rgt1.T == 0
plot3d(
    salt_mask.astype(np.float32),
    cmap='Reds',
    frames=[45, 73, 60],
    ifnewfig=True,
    figname="./gallery/salt_mask.png",
)

---

Plotting helpers in quick_start/plotting.py are adapted from pyseistr (https://github.com/aaspip/pyseistr) utilities.