SeisNetInsight

SeisNetInsight is an open-source Python package designed to help seismic network operators, researchers, and monitoring teams identify where new seismic stations would provide the greatest impact in improving earthquake location accuracy and depth resolution. The tool allows users to load their own earthquake catalog, stations, and optional contextual data (such as injection activity, faults, or population). Using these inputs, SeisNetInsight creates easy-to-interpret maps that highlight locations where additional stations would have the highest impact.

The method is based on well-established and widely accepted seismological principles describing how earthquake detection and location accuracy depend on the seismic network geometry, especially source-to-station distance and azimuthal coverage. These principles are incorporated as physically meaningful constraints within the automated workflow.

Package overview

The package exposes utilities for:

• Loading seismic event, station, and SWD well (optional) catalogues.
• Generating subject proximity (primary/secondary), ΔGap90, and SWD influence grids.
• Computing a composite monitoring index with user-defined weights and half-life decay.
• Exporting map artefacts (PNG, KMZ) from the computed grids.
• Launching an interactive web UI to guide users through data loading, grid computation, and map visualisation.

Installation

pip install .

This installs the seisnetinsight package along with the seisnetinsight-app console script for the Streamlit interface.

Running the web app

After installation, launch the interface with:

seisnetinsight-app

The UI is divided into three sections:

1. Data loading – Restore a previous session or configure a new one by uploading events, stations, optional SWD well catalogues, and BNA polygons. The interface shows the first 10 rows of each dataset, warns about missing expected columns, and optionally applies BallTree catalogue reduction.
2. Grids computation – Compute subject, ΔGap90, SWD, and composite grids with progress indicators, stop controls, and automatic session caching.
3. Maps – Explore interactive pydeck-based maps for each feature and the composite priority map. Download PNG and KMZ enabled maps.

GUI Overview:

Export the Maps to Google Earth

Programmatic usage

For more details see the examples in the notebook folder

from seisnetinsight.config import GridParameters
from seisnetinsight.data import load_events, load_stations, load_swd_wells
from seisnetinsight.grids import (
    compute_composite_index,
    compute_gap_grid,
    compute_subject_grids,
    compute_swd_grid,
    generate_grid,
    merge_grids,
)

bna_path = Path("../sample_files/midland.bna")
bna_bytes = bna_path.read_bytes()

SWD_COLUMN_MAP = {
    "latitude": "latitude_wgs84",
    "longitude": "longitude_wgs84",
    "volume": "SUM_injected_liquid_BBL",
}

PARAMS = GridParameters()

# events_df, stations_df, swd_df are pandas dataframes with latitude and longitude columns (more details in the notebooks folder)
grid = generate_grid(params)

subject_df = load_or_compute("subject",
    lambda: compute_subject_grids(events_df, stations_df, grid, params), overwrite=False)

gap_df = load_or_compute(
    "gap", lambda: compute_gap_grid(events_df, stations_df, grid, params), overwrite=False)

swd_grid = compute_swd_grid(swd_df, grid, params)
swd_grid = load_or_compute(
    "swd", lambda: compute_swd_grid(swd_df, grid, params), overwrite=False)

merged = merge_grids(subject_df, gap_df, swd_grid)
composite_grid_df = load_or_compute(
    "composite",
    lambda: compute_composite_index(
        merged,
        params),
    overwrite=False)

Output Examples

Panels (a–c) show how adding a new station would affect nearby events. Brighter colors mean that placing a station in that grid cell would impact more recent events. (a) shows how many recency-weighted events would fall within 4 km of a station. (b) shows the same for a 10 km radius. (c) shows how many events would have their maximum azimuthal gap reduced to ≤90° if a station were placed there. (d) shows the total salt-water-disposal volume within 25 km (in barrels). The black outline marks the Midland Basin:

(a) Map of a composite impact index (ranging from 0 to 1) that combines four metrics—S₄, S₁₀, ΔGap ≤ 90°, and SWD after adjusting for scaling, and weighting. Brighter colors highlight areas where a new station would have the most combined benefit. (b) Priority areas determined by k-means clustering of these metrics, ranked by their average composite index (Very High, High, Medium, Low). Blue triangles show current stations; the black outline marks the Midland Basin.