U.S. Wildfires from 2005 to 2015

Wildfire has always had a crucial role in the landscape of North America, making significant impacts on regional ecology, hydrology, and other components of the environment (“The Science Analysis”). Many ecosystems have become dependent on fire as a disturbance mechanism, requiring periodic fires to maintain species abundance and diversity (“Fire-Adapted”). More recently, wildfire has gained national attention due to its destructive capacity in regard to property and, more importantly, human life. It is estimated that the 2025 Los Angeles fires, killing 29 people, had an economic impact on the order of $250 billion (Irfan 2025). With an average of 70,000 wildfires occurring across the United States every year, it’s difficult to determine the total economic cost from these disturbances. The danger to human life, however, has resulted in an average of 12 deaths per year since 1980 (Smith 2025). Because of this, wildfire research has become an increasing priority, with the hope that a better understanding of past wildfires will lead to better management in the future.

Last summer, I got the opportunity to assist with fire science research through an internship with the U.S. Geological Survey. Our team surveyed vegetation recovery in the Black Hills of South Dakota and Wyoming, as well as Sequoia National Park.

Regeneration in a high mortality Sequoia grove	Me standing inside a Sequoia burn scar

This experience gave me the opportunity to see firsthand the aftermath of large-scale wildfires. As a native Virginian, I had no idea how important wildfire is to many ecosystems across our country or how destructive these fires can be when improperly managed. For instance, giant sequoias require fire for their seeds to germinate in dense undergrowth. But the KNP Complex Fire of 2021 grew to such proportions that it consumed many of these historic trees, some of which were likely centuries old. While my fieldwork gave me great insight into these two locales, I was interested in looking at wildfires on a national scale. I also found myself curious about ignition causes, both natural and anthropogenic. This project aims to investigate some of the basic characteristics of wildfires in the United States and how they relate to ignition causes, both over time and space. The data shown was originally sourced from the Fire Program Analysis fire-occurrence database, a collection of wildfire reports from several government agencies, covering the years 2005 to 2015. By looking at this period of time, I hoped to gain a better understanding of trends in wildfire number and size, as well as ignition causes, in the United States.

Figure 1 Figure 1 shows the total number of fire occurrences per year, colored by region. Contrary, perhaps, to popular belief and my own assumptions, the annual number of wildfires has not increased over time. If anything, the total number of wildfires per year may have slightly decreased from 2005 to 2015, with 2006 being an anomalously active year. As mentioned earlier, the annual average number of fires is 70,000, though with plenty of variability. I was surprised to see that most wildfires occur in the Southern United States, a region that I hadn’t strongly associated with fire previously. The Midwest and Northeast have comparatively fewer fires per year, which I suspect reflects their differing land use, climate, and plant communities. 2015 had a more evenly distributed year, with Southern states having the most fires but fewer than average, while the Midwest and Northeast had more than usual. However, the number of fires recorded in a region doesn’t give a complete picture of fire activity, as fires vary greatly in their size, just as states vary within their region. With this in mind, I wanted to get a better look at fire occurrences and burned area within each state.

Figure 2

There’s a clear trend between the number of fires and the total area burned in a state, with states featuring a greater total number of fires also generally having a greater burned area. This trend is visible throughout the entire time series, though with some interesting outliers. In 2015, Hawaii had only four recorded wildfires, and yet almost 5,000 acres were burned, which also coincides with a particularly active year for the Kīlauea volcano. California and Texas often appear as top states in both acres burned and number of fires, best seen in 2008, but also accompanied by other southern and western states. As mentioned earlier, there’s a clear relationship between region and fire occurrence, with the midwestern and northeastern states generally being on the lower ends of both axes. However, the lowest state, or more properly federal district, in both variables is consistently Washington, D.C. Seeing these relationships stay relatively consistent through time made me wonder if other variables experienced more change. Therefore, I investigated each of the categorical variables in my dataset.

Figure 3

• Cause — All causes appear to be declining over time, likely coinciding with the overall decrease in fire occurrence shown in Figure 1. Arson, in particular, seems to have significantly fallen over the years, going from the 3rd most common ignition cause to the 4th. Its closest competitor, lightning, appears to be following a similar trend, which wouldn’t necessarily be expected from a natural cause. The only increase can be found in structure fires, which also happen to be one of the least common causes.
• Landowner — it’s unclear whether the trends here are reflecting actual changes in fires by landowner over time or landowner recording practices, since the most obvious change is in the number of fires reported with missing and/or undefined. Fires on private property also increased sharply after 2008. Less noticeably, land managed by the Bureau of Indian Affairs, Bureau of Land Management, and US Forest Service has all experienced a decline in fire number over time.
• Size class — While most of the figure follows a general trend of fewer fires as size class increases, the second smallest class (B) is actually the dominating category over the entire time series. It also appears that this class of fires has gotten closer to size class A over time, begging the question of whether this trend continued after 2015. This data also highlights 2006 and 2011 as being particularly active years, as seen in Figure 1. It’s likely that climatic variables increased fire risk nationally, though with Texas and California particularly affected, as seen in Figure 2.
• Region — As previously discussed, the South clearly experiences more wildfires than other parts of the country. It’s also experiencing the sharpest drop in fire occurrence, though with great variability. Other regions appear to have less variability, though it also appears that the Northeast and Midwest went through a rise in fire occurrence from 2011 to 2015.

From these analyses, I found ignition cause to be the most compelling. Not only is understanding the cause of ignition a key part of prevention, but it also provides insight into what anthropogenic factors are driving wildfire outbreaks. The two fires I visited over the summer were ignited by arson and lightning, which made me wonder if geography had any part to play.

Figure 4

There definitely appears to be a spatial component, as the most common ignition cause varies substantially by region. Lightning is the leading cause of wildfires in the western half of the country, while the east shows much more variation. Interestingly, the Rust Belt and much of the Appalachian Mountains are dominated by debris burning, as well as our miscellaneous and undefined categories. Arson is prominent along the middle of the country, which happens to align with my own experience visiting a midwestern fire started by arson.

Structure being the least common makes sense with our earlier time series. But it’s also interesting to see that there’s more diversity in rarer causes. Again, there seems to be a spatial component, with fireworks being the least common cause in many eastern states but none west of the Mississippi River. Similarly, the Midwest is distinguished by power line- and railroad-caused fires as the rarest. I was also surprised to see that arson is the rarest cause in just one state, that being Delaware. After seeing that ignition showed so much geographic variation, I decided to look at whether seasonality had any effect.

Figure 5 This figure shows the mean proportion of fires by each ignition cause (in gray), as well as how that proportion changes by season. I’ll be honest, I figured that lightning would vary quite significantly by season. So it’s not surprising to see that most lightning ignitions occur during the summertime, while very few occur in winter. Debris burning, however, I did not expect to have the seasonal variation that it does. It also appears to have the opposite relationship, with more ignitions occurring during the winter. Similarly, arson doesn’t appear to be a summertime activity. Other, rarer causes don’t show much seasonality at all, such as structure, power lines, and smoking fires. The one exception seems to be fireworks, which cause many more fires during the summer than other seasons.

While some of these relationships may be explained by the natural sciences, like lightning causing more fires during the summer, others point towards variation in human behavior. Ignition by fireworks has an obvious connection to variable human activity, but it’s not clear when debris burning or arson would have such strong seasonal variation. These findings then led me to wonder if there was a connection between ignition cause and land ownership, as I figured that land ownership may reflect the level of human activity in an area.

Figure 6 Unfortunately, as seen earlier in Figure 3, the most predominant landowner in this dataset is Missing and/or Undefined. Therefore, the majority of the hotspots in this heatmap occur within the “Other” category, which includes those labeled “Foreign” and “Missing/Undefined.” However, there is a notable exception, with many fires caused by lightning being on federal land. I was also surprised to see that so many of the cold spots occur in the Tribal category. I suspect that this may just reflect the fact that there’s much less tribal land in the United States compared to these other groups. Arson appears to be one of the leading causes of wildfire on tribal lands, while railroad-caused fires are the rarest. Overall, I think this plot demonstrates a need for better information, as much of the data is hiding under the “Other” category.

It became clear to me through all of these visualizations that there is considerable variation in ignition cause across all of these variables, with unique relationships indicated in each. But when people think of wildfires, we tend to gravitate to those with the most news coverage, the ones that consume hundreds of thousands of acres over months. My own field experience was dominated by large-scale wildfires that got lots of press, as these events more easily justify their importance on grant proposals. Therefore, I decided to hone in on the largest fires by acreage in my dataset.

Figure 7 All of these fires fall into the class of megafires, which are defined as those that exceed 100,000 acres in size. Despite a large portion of fires being located in the southern U.S., as seen in Figures 1, 2, and 3, this map clearly shows that the majority of megafires occur in the western United States. There are a large number along the borders of Idaho, Oregon, and Nevada, as well as within the state of Alaska. I was also surprised to see that lightning was the predominant cause of these fires, though with notable exceptions. Some of the largest wildfires had other causes, such as a campfire in the case of the 2011 Wallow Fire of Arizona. The only megafires to have occurred during this time east of the Mississippi were located in southern Georgia and also caused by lightning.

Megafires are projected to increase in both frequency and size over the coming years due to compounding effects from climate change (i.e., rising temperatures, worsening droughts, and more lightning strikes) (Struzik 2020). Not only does this pose a threat to our nation’s forests and rangelands, it will likely take more human lives as well. While the western states will probably bear the brunt of this damage due to their arid climate and already fire-prone ecosystems, everyone will be affected. Throughout 2023 and 2024, smoke from Canadian wildfires sparked air quality advisories across the country (Zerkel 2023).

Wildfire smoke in Washington D.C., 2023

Thus, wildfires carry a health risk even to people thousands of miles away from the blaze itself (“Wildfires, Climate Change and Air Pollution” 2025). Megafires are particularly impactful due to their longer duration times, allowing them to expel pollutants for months on end.

Figure 8 Figure 8 shows the duration distribution of each size class, with the largest size classes (F and G) burning much longer than other, smaller fires. There are a few notable outliers, with a few class A fires reported as burning for several years despite never growing to more than an acre. Nevertheless, it’s clear that, generally speaking, larger fires burn longer, increasing their potential for both direct and indirect fatalities.

As I mentioned earlier, I’m a Virginian. Prior to my internship, I had never thought about wildfires one way or another, as they just hadn’t had much impact on my life. But through that experience and this project, I’ve gained a better understanding of fires at a local and national level, their causes, and their consequences. My hope is that through research such as this, we can make better management decisions in the future that both reduce the risk to human life and preserve our natural resources to the best of our ability.

References

“Fire-Adapted: Plants and Animals Rely on Wildfires for Resilient Ecosystems.” Defenders of Wildlife, defenders.org/blog/2020/07/fire-adapted-plants-and-animals-rely-wildfires-resilient-ecosystems. Accessed 19 Apr. 2025.

Irfan, Umair. “The LA Fires Have a Shocking Price Tag - and We’ll All Have to Pick up the Tab.” Vox, 3 Feb. 2025, www.vox.com/climate/397756/la-wildfire-insurance-palisades-california-fair-plan-climate.

“The Science Analysis of the National Cohesive Wildland Fire Management Strategy.” Vegetation and Fuels | The Science Analysis of The National Cohesive Wildland Fire Management Strategy, National Science Analysis Team, cohesivefire.nemac.org/vegetation-fuels. Accessed 19 Apr. 2025.

Smith, Adam B. “2024: An Active Year of U.S. Billion-Dollar Weather and Climate Disasters.” NOAA Climate.Gov, National Oceanic and Atmospheric Administration, 10 Jan. 2025, www.climate.gov/news-features/blogs/beyond-data/2024-active-year-us-billion-dollar-weather-and-climate-disasters.

Struzik, Ed. “The Age of Megafires: The World Hits a Climate Tipping Point.” Yale Environment360, Yale School of the Environment, 17 Sept. 2020, e360.yale.edu/features/the-age-of-megafires-the-world-hits-a-climate-tipping-point. “Wildfires, Climate Change and Air Pollution: A Vicious Cycle.” Clean Air Fund, 3 Apr. 2025, www.cleanairfund.org/news-item/wildfires-climate-change-and-air-pollution-a-vicious-cycle/.

Zerkel, Eric. “A New Outbreak of Canadian Wildfires Is Sending a Plume of Unhealthy Smoke into the US yet Again.” CNN, Cable News Network, 15 July 2023, www.cnn.com/2023/07/14/us/canada-wildfire-smoke-us-air-quality/index.html.

Image Credit

https://www.google.com/url?sa=i&url=https%3A%2F%2Fwww.politico.com%2Fnews%2F2023%2F06%2F29%2Fair-quality-washington-alert-00104161&psig=AOvVaw3Ur3uvhHbpb0CooW8oevBJ&ust=1745154880271000&source=images&cd=vfe&opi=89978449&ved=0CBQQjRxqFwoTCNiJnteW5IwDFQAAAAAdAAAAABBM