It is hard to ignore the fact that wildfires are getting worse across North America. Even those who haven’t seen them up close or lost their homes have probably inhaled the smoke. This report shares new findings from Deep Sky Research and its Wildfire Risk Model on how wildfire risk is changing due to climate change.

‍

Extreme Fire Weather

One way to understand the increase in wildfire destruction is to look at the underlying conditions that lead to wildfires. Fire Weather Index (FWI) is a measure developed by the Canadian Forestry Service but used globally to assess fire risk. It combines temperature, humidity, wind speed, and precipitation to give a holistic view of landscape flammability. 

Climate change, which directly impacts each of these inputs, is causing more frequent and more destructive wildfires. At Deep Sky Research we analyzed trends in fire weather across North America and found sharp increases in the probability of extreme conditions. 

The map below shows how fire weather is changing in each county of the continental US. It shows how Extreme Fire Weather is becoming more frequent and more severe. The darker red the county the greater the increase in fire weather. The few blue counties are actually seeing decreases in risk.

‍

‍

In our post on climate models we advocate for more insurance-style risk assessment in order to better understand climate risk. We have applied this approach in our wildfire research. Before issuing insurance for your home or car, actuaries examine risk on two levels: frequency and severity. How often should we expect a bad outcome and how bad could it be? On both those levels, Extreme Fire Weather is getting worse. The maximum potential fire weather across North America has increased sharply, and the frequency of Extreme Fire Weather has also shot up.

At Deep Sky Research we built a Wildfire Risk Model to explain how wildfire risk is changing due to climate change. The model provides several important findings.

‍

1: Maximum Fire Risk Has Grown Approximately 15X Across North America

The model shows that Extreme Fire Weather conditions previously seen once every 100 years will now happen on average every 7 years. When ignition happens – due to lightning, arson, or the spark from a train – wildfires will grow bigger, faster and cause more damage to property and human life.

The most destructive and deadly fires in history were fueled by Extreme Fire Weather conditions. It is clear from their track record that fire weather having a higher and higher ceiling is extremely dangerous. More information below on the importance of the Extreme Fire Weather metric. 

‍

2: Frequency of Extreme Fire Risk has Grown 20X 

The frequency of Extreme Fire Weather – exceeding the 95th percentile in that area – has grown even more sharply than severity. Levels of widespread extreme fire risk which used to occur once every 100 years will now occur every 5. We will see more Extreme Fire Weather throughout the year with wildfire seasons starting earlier in the Spring and ending later in the Fall. In some parts of the country wildfires are raging year-round. 

A change in the fire weather conditions is often what helps reign in an out of control wildfire. Rain or a drop in temperature can cause a wildfire to naturally subside. More fire weather means these potentially life-saving changes in weather are growing less and less common. The odds that tomorrow will also have Extreme Fire Weather are growing. 

‍

3: Some Regions Face Even Faster Growing Risks

Click through the slides below to see where Extreme Fire Weather is growing fastest and the one region where it’s decreasing. 

‍

‍

4: The Increase is Accelerating

One startling finding from the Wildfire Risk Model is that the increases in Extreme Fire Weather are not linear. These risks are not only growing but have begun accelerating. Non-linear trends are challenging to grasp, especially when studying climate data that represents a relatively long period of time. Climate models which observed a specific relationship between variables in the past may already be out of date if that relationship is changing over time.  

For this reason, the Wildfire Risk Model allows for the relationship between Extreme Fire Weather and time to change as time progresses and the planet warms as we continue to emit. 

‍

Why Focus on “Extreme Fire Weather”?

Fire weather does not capture all important aspects of wildfires. It ignores sources of ignition for example. But research has shown that extreme fire weather is strongly predictive of the severe wildfires which are to blame for the mass evacuations and destruction we’ve seen in recent years. These are fires that are simply beyond our ability to control. They include the Fort McMurray fire of 2016 in Alberta, Canada, the 2018 Camp Fire which killed 85 people and caused over 16 B USD in damages, and the wildfires across Canada last summer which burned 45 million acres of forest. 

The visualization below shows how four of the most destructive fires in North American recorded history stack up on fire weather. Each dot represents a year, placed along the x axis by the percentage of Extreme Fire Weather in that county during that period. The dots in each row are filtered for the specific months of the fire of interest to give an accurate comparison.

‍

‍

Each of these fires are at or very near the extreme right edge of the county’s scale. This demonstrates the importance of analyzing trends in Extreme Fire Weather. We will not be prepared for fires as destructive as these if we don’t. 

‍

Wildfire Emissions

The wildfires in Canada last summer had devastating impacts, from the 45 million acres burned to the hundreds of thousands who had to evacuate their homes, and the air pollution which spread across the entire continent. Another impact is troubling in the longer term: the carbon emissions from these fires. Carbon emissions from wildfires in Canada in 2023 were more than double those from the oil and gas sector, the country’s largest emitter. 

‍

‍

This is what climate scientists refer to as a “vicious cycle.” Climate change is causing worse fire weather conditions, which cause larger, more destructive wildfires, which cause huge carbon emissions, which themselves contribute to more climate change. Vicious cycles like this one are deeply concerning. If the death and destruction caused by wildfires is not reason enough to act on climate change, avoiding vicious cycles and tipping points surely is. 

‍

Methodology

The fundamentals of Deep Sky’s research are laid out in this post on how we build our models. This section contains some additional details on the analysis in this report. 

This report analyzed over 40 years of fire weather data from 1973-2024 across North America. The data was made available by research from Vitolo et al. 2020. From the FWI data two metrics were constructed: Extreme Fire Weather severity and Extreme Fire Weather frequency. Severity measures the annual expected maximum FWI in a given location and frequency consists of the number or percentage of occurrences of FWI greater than the 95th percentile for that location. Baseline period, unless otherwise specified, refers to the period from 1973 until 1999. 

This report used extreme value statistics as described here to model trends in Extreme Fire Weather and calculate return periods. This paper uses a comparable methodological approach and found similar results in Europe. A Technical Appendix is available upon request.