GIS

Determining the Susceptibility of Landslides within the Thomas Fire Boundary

teodulfo.espero

July 30, 2025

Introduction

When the Earth decides to shift, slide, and tumble, we get what the USGS politely calls a “landslide”—a down-slope movement of rock, debris, or earth. Triggered by gravity and exacerbated by factors like rainfall, earthquakes, and yes, human meddling, landslides are the ultimate uninvited guests in California’s rugged terrain. One major culprit behind increased landslide risks? Wildfires. And in Southern California, the 2017 Thomas Fire proved just that.

This project uses Geographic Information Systems (GIS) to determine where landslides are most likely to occur within the burn scar of the Thomas Fire.

Background

The Thomas Fire consumed over 281,000 acres across Ventura and Santa Barbara Counties, burning fiercely from December 4, 2017, to its eventual containment in early January 2018. The aftermath wasn’t just scorched earth and lost homes—it primed the region for landslides.

Using spatial data, satellite imagery, rainfall records, terrain elevation models, and QGIS processing tools, we can map out high-risk landslide zones. Because what better way to fight natural disasters than with pixels and algorithms?

Methods

Image Files and Satellite Data

We sourced Landsat 8 imagery (specifically Bands 3 and 4) with <10% cloud cover from the USGS EarthExplorer. Vegetation loss was notable within fire-impacted zones, making this a key landslide indicator.

Landsat 8 Breakdown:

  • Band 3 (Green): Shows vegetation
  • Band 4 (Red): Shows man-made vs natural features
  • Spatial resolution: 30m

Rainfall Data via NASA Giovanni

Using NASA’s Giovanni tool, we downloaded precipitation data for January 1–10, 2018. Rainfall increases landslide risk, and Giovanni’s IMERG data made it possible to visualize the cumulative rainfall during that critical window.

Fire Perimeter and Terrain

We obtained the Thomas Fire boundary shapefile from USGS’s GeoMAC dataset and terrain data from the GMTED2010 elevation model. With this, we processed slope steepness using QGIS’s GDAL Slope tool, classifying gradients up to 90°.

Weighted Overlay and Classification

We calculated a Landslide Susceptibility Index (LSI) using:

  • NDVI (vegetation index)
  • Slope (degree of incline)
  • Rainfall (accumulated)

Each factor was reclassified into five index values, then weighted using:

  • Vegetation (weight: 2)
  • Slope (weight: 1)
  • Rainfall (weight: 1)

The formula:

plaintextCopyEditLSI = (2 × NDVI) + (1 × Slope) + (1 × Rainfall)

This resulted in a final raster showing landslide-prone zones—the brighter the color, the higher the risk.

Results and Analysis

Our processed map revealed extensive high-risk zones within the Thomas Fire perimeter, especially in areas with steep slopes and minimal vegetation. A Kappa Index of 0.8423 and overall accuracy of 91% means our classification was strong. Roads like Highway 101 showed clear vulnerability, highlighting risks to infrastructure.

Despite some vegetation regrowth, the land remains vulnerable. Burnt areas with thin or no vegetation are still on the danger list, regardless of how green they may seem today.

Conclusions

GIS proves to be a powerful ally in post-wildfire disaster risk assessment. With open-access datasets and QGIS, we identified landslide-prone zones that can inform urban planning, evacuation strategies, and reforestation priorities.

While this project was limited to 2018 imagery and available datasets, future work could incorporate:

  • Soil type and geology
  • Specific vegetation types
  • More refined elevation data

In sum, with fire comes vulnerability—but with GIS comes visibility.

The PDF copy of the study can be seen here.

References

Tags: