Scorched Earth: A Journalist’s Guide to Investigating Wildfires with Open Source Intelligence

In an era defined by a rapidly shifting climate, wildfires have moved from seasonal occurrences to persistent, systemic threats. For researchers, journalists, and environmental advocates, detecting a fire is only the beginning. While tools like NASA’s FIRMS provide immediate heat signatures and coordinates, they offer little insight into the long-term ecological damage, the specific severity of the destruction, or the patterns of recurrence that suggest a region is becoming increasingly vulnerable.

This guide provides a comprehensive methodology for investigating the scale and severity of wildfires, using the 2025 blaze at Sicily’s Zingaro Nature Reserve as a primary case study. By integrating satellite imagery from the European Space Agency’s Copernicus program with spatial analysis in QGIS, we can transform raw data into actionable findings.

The Main Facts: Defining the Scope of Destruction

When investigating a wildfire, the primary objective is to establish a verified timeline and spatial footprint. The Capo San Vito peninsula in Sicily, home to the ecologically fragile Zingaro Nature Reserve, serves as a sobering example of how quickly fire can devastate protected lands. In July 2025, a massive wildfire swept through the reserve, destroying ancient forests, vital grasslands, and agricultural plots.

The impact was so profound that sections of the reserve remain closed to the public today. To report on such an event, investigators must move beyond anecdote. By utilizing data from the Landsat and Sentinel-2 missions, we can generate a "before and after" snapshot that captures the fire’s progression with scientific precision.

Visualizing the Invisible

While true-colour satellite imagery—which mimics the human eye—shows general changes in the landscape, it often fails to highlight the full extent of scorched earth. To overcome this, researchers use multispectral imaging. Sentinel-2 captures wavelengths outside the visible spectrum, specifically Near-Infrared (NIR) and Shortwave Infrared (SWIR).

By assigning these bands to specific colour channels (NIR as green and SWIR as red), we create a "false-colour" composite. In this view, healthy, chlorophyll-rich vegetation appears vibrant green, while burned, carbonized earth appears a stark, unmistakable red. This technique allowed us to determine that the fire ravaged the Capo San Vito peninsula between July 20 and July 27, 2025.

Chronology: Mapping the Timeline of a Disaster

Establishing a timeline is essential for accountability. By browsing the Sentinel-2 catalogue via the Copernicus Browser, investigators can identify the last available "healthy" image before the ignition and the first clear, cloud-free image after the fire was extinguished.

1. Pre-Fire Baseline: Access imagery from mid-July 2025 to establish the density and health of the vegetation.
2. Detection Window: Cross-reference satellite timestamps with local reports to pinpoint the onset of the fire.
3. Post-Fire Assessment: Utilize the earliest post-event imagery to define the final burn perimeter.

By drawing an "Area of Interest" (AOI) polygon within the Copernicus Browser, we calculated that the 2025 blaze consumed over 50 square kilometres of the peninsula. This quantitative approach elevates reporting from qualitative descriptions of "a large fire" to verified findings based on remote sensing data.

Supporting Data: Quantifying Severity via NBR

Not all fires are created equal. A fire that consumes only surface litter may allow an ecosystem to recover, while a high-severity blaze can sterilize soil and prevent reforestation for decades. To measure this, we use the Normalised Burn Ratio (NBR).

The NBR relies on the spectral difference between NIR (which is highly reflective in healthy plants) and SWIR (which is highly reflective in burned, dry, or exposed soil). The formula is:

NBR = (NIR – SWIR) / (NIR + SWIR)

A high NBR value (closer to 1) indicates dense, healthy vegetation, while a low or negative value indicates a high degree of burn. By applying custom scripts within the Copernicus Browser or calculating these values in QGIS, we found that the July 2025 fire was "moderately severe" according to U.S. Forest Service classification guidelines.

Official Responses and Ecological Implications

The destruction of protected areas like the Zingaro Nature Reserve poses a critical question: how do these fires affect areas of high conservation value? By importing the burned area raster from Copernicus into QGIS—the industry-standard open-source geographic information system—and overlaying it with the Protected Planet dataset, we can visualize the overlap between the blaze and the reserve’s boundaries.

Using the Zonal Histogram tool in QGIS, we performed a pixel-count analysis. The results were stark: approximately 96.3% of the Zingaro Nature Reserve was impacted by the fire. This finding illustrates the failure of current land management practices to prevent catastrophic damage in zones specifically designated for ecological protection.

The Human and Climate Factor

While natural factors like drought and extreme heat contribute to these events, land management—or the lack thereof—is a recurring theme in European wildfire reports. As noted by the European Commission, the EU has had to deploy record-breaking wildfire responses in recent years, signaling that current preventative measures are struggling to keep pace with the climate-induced acceleration of fire seasons.

Historical Context: Are Fires Becoming More Frequent?

To understand if the 2025 fire was an anomaly or part of a trend, we must look at historical data. The European Forest Fire Information System (EFFIS) provides an invaluable archive of fire incidents dating back to 2008.

By downloading the EFFIS wildfire data for Italy (2015–2025) and performing a spatial intersection in QGIS, we can isolate only the fires that occurred within the Zingaro reserve boundaries. Using the Data Plotly plugin to visualize this data, a clear narrative emerges: the reserve has been hit by multiple significant fires over the last decade. However, the 2025 event was larger in scale and more destructive than the major fires of 2017 and 2020.

This longitudinal analysis is crucial for policy discussions. It suggests that the peninsula is not just experiencing random events, but is entering a cycle of increased vulnerability, where the recovery time between fires is shrinking—a hallmark of climate-stressed ecosystems.

Conclusion: Empowering Future Investigations

The tools detailed in this guide—Copernicus Browser, QGIS, and the EFFIS database—are free, accessible, and powerful. They allow any investigator to move beyond the superficial "hotspot" reports and into the realm of rigorous, evidence-based environmental journalism.

By following these steps, you can:

• Verify the physical extent of a fire using multispectral imaging.
• Quantify the damage using the Normalized Burn Ratio.
• Contextualize the event by mapping it against historical trends.

Whether you are investigating a wildfire in Sicily, the forests of the Amazon, or the grasslands of Chile, the methodology remains the same. As climate change reshapes the global landscape, the ability to independently verify and report on the scale of ecological loss is not just a skill—it is a responsibility.

For those looking to deepen their research, Bellingcat’s toolkit on QGIS and previous guides on NASA FIRMS serve as essential resources for building a robust investigative workflow.