Comparison of the Eaton and Palisades Fires
Author: William Mullins ・ Date: December 12, 2025
View Additional Content on GithubAbout
On January 7, 2025, the Eaton and Palisades fires ignited nearly simultaneously in the Los Angeles metropolitan area, devastating thousands of acres and displacing countless residents. These fires became two of the most destructive wildfires in California history, prompting questions about the environmental and demographic impacts they may have had.
This analysis uses satellite remote sensing and census data to accomplish two objectives:
- To visualize the burn scars using false color imagery that can penetrate the smoke and highlight fire damage
- To examine if elderly populations were disproportionately affected by comparing data from within each fire's perimeter
Highlights
- Geospatial data wrangling using geopandas for vector operations and xarray with rioxarray for raster processing
- False-color composite visualization combining SWIR, NIR, and Red bands to reveal burn scars invisible in natural color imagery
- Spatial analysis using census tract information and fire perimeters to quantify population impacts
Datasets
A geodatabase created by Centers for Disease Control and Prevention and Agency for Toxic Substances Disease Registry concerning environmental justice at the county level in the United States. A subset of this data concerning counties in California in 2024 was used in this analysis.
GeoJSON files were downloaded containing the perimeter data for the Palisades and Eaton fires.
Landsat 8-9 data collected by NASA showing surface reflectance and temperature. The data used in this analysis is a NetCDF containing a subset of this data which centers on the area surrounding the two fires.
Analysis Setup
Loading Required Packages
False Color Image Analysis
False-color composites are widely used to assess wildfire impacts. By combining non-visible spectral bands—such as short-wave infrared (SWIR) and near-infrared (NIR)—these images can penetrate smoke and reveal burn scars that aren't apparent in natural-color imagery.
Reading Landsat Satellite Data
The satellite data, stored in a NetCDF file, is loaded using xarray.open_dataset(). The coordinate reference system (CRS) is then restored with rioxarray.
Generating the False Color Composite
This code builds a false-color composite by selecting three spectral bands and assigning them to the RGB channels: SWIR to red, NIR to green, and red (visible light) to blue. This combination makes burned areas appear bright red-orange, while healthy vegetation appears green and urban areas appear gray or tan. Those values are then scaled using robust scaling, which reduces the impact of outliers, and plotted to give the map below.
Loading Fire Perimeter Data
The fire perimeter boundaries are read from GeoJSON files and reprojected to match the Landsat data's CRS.
Overlaying Fire Perimeters on False Color Composite
The fire perimeters are overlayed on the false color composite to better show the range of the fire. From this it can be seen that both of the regions have significant scarring, with the color in the Palisades region suggesting more severe burning.
Age Disparity Analysis
Beyond visualizing damage, it's important to understand the populations affected by the fires. This section examines whether elderly populations, who may experience greater challenges during evacuation and recovery, were disproportionately impacted by either fire.
Loading Environmental Justice Index Data
The CDC's Environmental Justice Index provides census tract–level demographic data, including percentile rankings for the population aged 65 and older. This dataset was read in and reprojected to match the boundaries CRS.
Creating Census Tracts from Data
The census data is then spatially joined with the fire perimeters to identify all census tracts overlapping each burned area. The resulting dataset includes only those intersecting tracts, along with their associated demographic attributes.
Visualizing Age Demographics by Fire Region
The final visualization displays the census tracts affected by each fire, colored by their EPL_AGE65 percentile value.
Interpretation
The results reveal a notable demographic disparity between the two fire-affected areas. Census tracts within the Palisades fire perimeter have a higher proportion of residents aged 65 and older than most tracts in California. In contrast, tracts impacted by the Eaton fire show a lower proportion of older adults relative to the state average. Future studies should examine whether this disparity influenced emergency response planning and resource allocation.
References
Centers for Disease Control and Prevention and Agency for Toxic Substances Disease Registry. [2024] Environmental Justice Index. Accessed [12/12/2025].https://atsdr.cdc.gov/place-health/php/eji/eji-data-download.html
Earth Resources Observation and Science (EROS) Center. (2020). Landsat 8-9 Operational Land Imager / Thermal Infrared Sensor Level-2, Collection 2 [dataset]. U.S. Geological Survey.https://doi.org/10.5066/P9OGBGM6
Palisades and Eaton Dissolved Fire Perimeters as of 2025/01/21. (2025). ArcGIS REST Services Directory.https://services.arcgis.com/RmCCgQtiZLDCtblq/arcgis/rest/services/Palisades_and_Eaton_Dissolved_Fire_Perimeters_as_of_20250121/FeatureServer