SHALSTAB - Shallow Landsliding STABility Model

A comprehensive Python package for slope stability analysis using the SHALSTAB (Shallow Landsliding STABility) model. This implementation provides tools for analyzing slope stability using physically-based models that consider topographic, hydrologic, and geotechnical factors.

Table of Contents

• Overview
• Installation
• Quick Start
• Scientific Background
• Mathematical Framework
• Model Components
• Usage Examples
• Contributing
• License
• References

Overview

The SHALSTAB model evaluates infinite slope stability by combining:

• Topographic analysis (slope, flow accumulation)
• Hydrologic modeling (steady-state water flow)
• Geotechnical parameters (cohesion, friction angle, unit weight, permeability)

Key Features

• Complete SHALSTAB implementation with all core analyses
• Raster-based processing using xarray and rasterio
• Professional visualizations with matplotlib
• Comprehensive error handling and validation
• Type hints and detailed documentation
• Export capabilities to GeoTIFF and other formats

Installation

Using pip (when available)

pip install shalstab

From source

git clone https://github.com/federicogmz/shalstab.git
cd shalstab
pip install -e .

Dependencies

pip install -r requirements.txt

Training data

The easiest way to test if SHALSTAB is installed correctly or to get familiar with its implementation is to use the provided training data:

import shalstab

# Use training data
analyzer = shalstab.Analyzer(
    dem_path=shalstab.training_dem,
    geo=shalstab.training_geology,
    geo_columns=["Cohesion", "Phi", "Gamma_kN_m", "Ks_m_s"]
)

# Run analysis
critical_rain = analyzer.calculate_critical_rainfall()

Scientific Background

SHALSTAB Model Overview

SHALSTAB (Shallow Landsliding STABility) is a physically-based model developed by Montgomery and Dietrich (1994) for analyzing shallow landslide susceptibility. The model combines:

1. Infinite slope stability analysis - Classical geotechnical stability assessment
2. Steady-state hydrologic model - Water flow and saturation analysis
3. Topographic controls - Digital terrain analysis for slope and drainage

The model is particularly suited for analyzing rainfall-triggered shallow landslides in natural slopes where the failure surface is parallel to the ground surface.

Model Assumptions

The SHALSTAB model is based on several key assumptions:

1. Infinite slope geometry: The slope extends infinitely in all directions, allowing for 1D analysis
2. Planar failure surface: Failure occurs along a plane parallel to the ground surface
3. Steady-state hydrology: Groundwater flow reaches equilibrium during rainfall events
4. Saturated conditions: The model considers fully saturated soil conditions at failure
5. Uniform soil properties: Geotechnical parameters are assumed constant within each geological unit
6. Coulomb failure criterion: Soil strength follows the Mohr-Coulomb failure envelope

Applications and Limitations

Suitable for:

• Regional-scale landslide susceptibility mapping
• Rainfall threshold analysis
• Comparative stability assessment
• Risk prioritization and hazard zoning

Limitations:

• Not suitable for deep-seated landslides
• Cannot predict exact timing of failure
• Requires high-quality input data
• May not capture complex hydrogeological conditions

Mathematical Framework

Core SHALSTAB Equation

The fundamental SHALSTAB stability criterion compares the ratio of upslope contributing area to contour length (destabilizing force) with a stabilizing term that includes soil properties and rainfall:

a/b ≤ (T/q) × [(c/(γw × z × cos²θ × tanφ)) + (γ/γw) × (1 - tanθ/tanφ)]

Where:

• a = upslope contributing area (m²)
• b = contour length (cell width, m)
• T = transmissivity (k × z × cosθ, m²/day)
• q = recharge rate (rainfall intensity, mm/day converted to m/day)
• c = soil cohesion (kPa)
• γw = unit weight of water (9.81 kN/m³)
• z = soil thickness (m)
• θ = slope angle (radians)
• φ = internal friction angle (radians)
• γ = soil unit weight (kN/m³)
• k = saturated hydraulic conductivity (m/s)

Critical Rainfall Calculation

The critical rainfall represents the minimum rainfall intensity required to trigger slope failure:

qcrit = (T/a) × [(c/(γw × z × cos²θ × tanφ)) + (γ/γw) × (1 - tanθ/tanφ)]

This equation is derived by rearranging the stability criterion to solve for the critical recharge rate.

Stability Classification

The model classifies each cell into one of four stability categories:

1. Unconditionally Stable (Class 1): tanθ/tanφ < γ/γw - Stable regardless of rainfall
2. Unconditionally Unstable (Class 2): tanθ/tanφ > 1 - Unstable regardless of rainfall
3. Unstable (Class 3): Left side > Right side for given rainfall - Unstable at specified rainfall
4. Stable (Class 4): Left side ≤ Right side for given rainfall - Stable at specified rainfall

Log(q/T) Analysis

The log(q/T) ratio provides a dimensionless measure of stability:

log(q/T) = log10(q × cell_size / flow_accumulated) - log10(T)

Where:

• Negative values indicate potentially unstable conditions
• Positive values indicate more stable conditions
• Values near zero represent marginal stability

Model Components

1. Soil Thickness Estimation

The package implements the Catani et al. (2010) empirical model for soil thickness estimation:

z = zmax × [1 - ((tanθ - tanθmin) / (tanθmax - tanθmin)) × (1 - zmin/zmax)]

Where:

• z = calculated soil thickness (m)
• zmax = maximum soil thickness (typically 5.0 m)
• zmin = minimum soil thickness (typically 0.1 m)
• θ = local slope angle (radians)
• θmax, θmin = maximum and minimum slope angles in study area

Model Assumptions:

• Linear relationship between slope and soil thickness
• Soil loss increases with slope steepness
• Minimum thickness maintained even on steepest slopes
• Maximum thickness occurs on flattest areas

Applications:

• Weight of soil column calculation (driving force)
• Available cohesive strength estimation
• Transmissivity calculations (combined with permeability)

2. Topographic Analysis

Slope Calculation

Slope angles are calculated using finite difference approximation:

slope = arctan(√((dz/dx)² + (dz/dy)²))

The calculation process involves:

1. Computing elevation gradients using finite differences
2. Calculating slope magnitude from gradient components
3. Converting from radians to degrees for display

Flow Direction and Accumulation

The package uses the D8 flow direction algorithm:

1. Flow Direction: Determines steepest descent direction for each cell
2. Flow Accumulation: Calculates upslope contributing area using D8 routing
3. Preprocessing: Fills pits and resolves flat areas for continuous flow routing

3. Hydrologic Modeling

Transmissivity Calculation

Soil transmissivity represents the ability to transmit water laterally:

T = k × z × cosθ

Where:

• T = transmissivity (m²/day)
• k = saturated hydraulic conductivity (m/s, converted to m/day)
• z = soil thickness (m)
• θ = slope angle (radians)

Steady-State Flow Model

The model assumes steady-state conditions where:

• Recharge rate equals rainfall intensity
• Groundwater flow reaches equilibrium
• Saturated zone thickness remains constant during analysis

4. Geotechnical Framework

Mohr-Coulomb Failure Criterion

Soil strength follows the Mohr-Coulomb relationship:

τ = c + σn × tanφ

Where:

• τ = shear strength (kPa)
• c = cohesion (kPa)
• σn = normal stress (kPa)
• φ = internal friction angle (degrees)

Infinite Slope Analysis

For infinite slope conditions, the factor of safety is:

FS = (c + (γ×z×cosθ - γw×zw×cosθ) × tanφ) / (γ×z×sinθ×cosθ)

Where:

• FS = factor of safety
• zw = height of water table above failure plane (m)

5. Failure Probability Assessment

The failure probability calculation provides a relative ranking of landslide susceptibility:

Process:

1. Calculate log(q/T) using calculate_log_qt()
2. Invert values: -log(q/T)
3. Normalize to 0-100% range: P = ((inverted - min) / (max - min)) × 100

Interpretation:

• Areas requiring high rainfall to fail → Low probability
• Areas requiring low rainfall to fail → High probability
• Unconditionally stable/unstable areas → No probability (NaN)

Important Note: The probability is relative, not absolute. It represents comparative susceptibility across the study area rather than actual failure rates.

Usage Examples

Basic Topographic Analysis

import shalstab

# Example 1: Using file paths (recommended)
analyzer = shalstab.Analyzer(
    dem_path="elevation.tif",
    geo="geology.geojson",
    geo_columns=["cohesion", "friction", "gamma", "permeability"]
)

# Example 2: Using GeoDataFrame
import geopandas as gpd
geology = gpd.read_file("geology.shp")
analyzer = shalstab.Analyzer("elevation.tif", geology, geo_columns=["cohesion", "friction", "gamma", "permeability"])

# Initialize full analyzer with file paths
analyzer = shalstab.Analyzer(
    dem_path="high_res_dem.tif",
    geo="geology_with_properties.geojson",
    geo_columns=[
        "cohesion_kpa",      # Soil cohesion (kPa)
        "friction_deg",      # Internal friction angle (degrees)
        "gamma_knm3",        # Soil unit weight (kN/m³)
        "k_ms"               # Hydraulic conductivity (m/s)
    ],
    figsize=(15, 10)
)

# Calculate critical rainfall with visualization
critical_rainfall = analyzer.calculate_critical_rainfall(show_plot=True)
print(f"Critical rainfall range: {critical_rainfall.min():.1f} - {critical_rainfall.max():.1f} mm/day")

# Analyze stability for specific rainfall events
rainfall_events = [10, 25, 50, 100]  # mm/day

for rainfall in rainfall_events:
    stability, fig = analyzer.calculate_stability(rainfall_mm_day=rainfall)

    # Calculate unstable area
    unstable_cells = (stability == 3).sum()
    cell_area = abs(analyzer.dem.rio.resolution()[0] * analyzer.dem.rio.resolution()[1])
    unstable_area_km2 = unstable_cells * cell_area / 1e6

    print(f"Rainfall {rainfall} mm/day: {unstable_area_km2:.2f} km² unstable")

    # Save results
    analyzer.export_raster(stability, f"stability_{rainfall}mm.tif")
    fig.savefig(f"stability_plot_{rainfall}mm.png", dpi=300, bbox_inches='tight')

# Calculate relative failure probability
probability = analyzer.calculate_failure_probability()
print(f"High risk areas (>80%): {(probability > 80).sum()} cells")

# Export all results
analyzer.export_raster(critical_rainfall, "critical_rainfall.tif")
analyzer.export_raster(probability, "failure_probability.tif")
analyzer.export_raster(analyzer.soil_thickness, "soil_thickness.tif")

Advanced Analysis Workflow

# Multi-scenario analysis
scenarios = {
    "dry_season": 15,      # mm/day
    "normal_rain": 35,     # mm/day
    "heavy_rain": 75,      # mm/day
    "extreme_event": 150   # mm/day
}

results = {}
for scenario, rainfall in scenarios.items():
    stability, fig = analyzer.calculate_stability(rainfall_mm_day=rainfall)

    # Calculate statistics
    stats = {
        'total_cells': stability.count().item(),
        'stable_cells': (stability == 4).sum().item(),
        'unstable_cells': (stability == 3).sum().item(),
        'uncond_stable': (stability == 1).sum().item(),
        'uncond_unstable': (stability == 2).sum().item()
    }

    # Calculate percentages
    total = stats['total_cells']
    stats['stable_pct'] = stats['stable_cells'] / total * 100
    stats['unstable_pct'] = stats['unstable_cells'] / total * 100

    results[scenario] = stats

    # Export scenario
    analyzer.export_raster(stability, f"stability_{scenario}.tif")
    fig.savefig(f"stability_{scenario}.png", dpi=300, bbox_inches='tight')

# Print summary
print("\nStability Analysis Summary:")
print("-" * 60)
for scenario, stats in results.items():
    print(f"{scenario:15} | Stable: {stats['stable_pct']:5.1f}% | Unstable: {stats['unstable_pct']:5.1f}%")

Data Preprocessing

# Preprocess DEM to fill NoData values
processed_dem = analyzer.preprocess_dem("raw_dem_with_gaps.tif")
print(f"Processed DEM shape: {processed_dem.shape}")

# Resample raster to different resolution
coarse_dem = analyzer.resample_raster("high_res_dem.tif", resolution=30.0)  # 30m resolution
analyzer.export_raster(coarse_dem, "dem_30m.tif")

Contributing

Contributions are welcome! Please see CONTRIBUTING.md for guidelines on how to contribute to this project.

License

GNU General Public License v3.0 (GPL-3.0)

See LICENSE for full license text.

References

• Montgomery, D.R., & Dietrich, W.E. (1994). A physically based model for the topographic control on shallow landsliding. Water Resources Research, 30(4), 1153–1171.
• Dietrich, W.E., & Montgomery, D.R. (1998). SHALSTAB: A digital terrain model for mapping shallow landslide potential. NCASI Technical Bulletin 796.
• Catani, F., Segoni, S., & Falorni, G. (2010). An empirical geomorphology-based approach to the spatial prediction of soil thickness at catchment scale. Water Resources Research, 46(5).