Research Triangle Energy Consortium

We now have the smoking gun

Eight of the ten largest wildfires ever in California occurred after 2017. The August Fire in 2020 burnt over a million acres. At the time, the next largest had been the Mendocino Complex in 2018, which burned 459000 acres. Then, the Dixie Fire in 2021was nearly as large as the record holder. Most, not all, scientists attribute this to climate change. But causality has been hard to establish. Then came the recent paper in The Proceedings of the National Academy of Sciences (PNAS) which was as close to a smoking gun as one could have when dealing with multiple variables. More on that later, but first some basics.

The severity of a fire season is often judged by the acreage burned, not the number of fires. This makes sense because that is the metric which is connected to the impact on society. The difference stands out when fire activity is examined over the last few decades. Shown in Figure 1 are US Environmental Protection Agency (EPA) wildfire figures for both metrics over three decades.

Courtesy US EPA 2019 (Wildland Fire Research Framework 2019-2022)

Figure 1. Time series of numbers of fires and area burned in the period 1988 – 2018.

The numbers of fires have no perceptible trend, although one can see a general progression downwards in the last decade. The acreage burned, on the other hand, has seen a gradual slope up, even though a regression line would show significant uncertainty due to the severe annual swings. Interestingly, the 2020 figure (not shown) is not that different from the 2017, even though five of the 10 largest fires ever were in 2020. A fire is considered large when the area burned is greater than 400 hectares (a hectare is roughly 2.47 acres). Acreage burned is dominated by the larger fires. Obviously, the impact on society is proportional to the area burned and the durations of the fires. One study showed that durations of large fire (>400 hectare) burns averaged 6 days in the period 1973-82 and was over 50 in 2003-12.

In attempting to unravel the distinction between the patterns in numbers and extents of fires, one statistic stands out. In recent years, ninety five percent of fires in California and the Mediterranean region have been caused by human activity (Syphard and Keeley, 2015). These activities include campfires, arson, equipment (such as chainsaws and mowers), vehicles, falling power lines, and controlled burns gone, well, uncontrolled. While human behavior can be influenced (the Smokey Bear program, for example), not surprising is the fact that the trend in the number of ignitions has been generally flat. Among natural ignition events lightning plays an interesting part. Depending on region, it ordinarily clocks in as the 4^th to 8^th most frequent contributor to ignition. The 2020 fire season changed all that. Cal Fire, a state agency reports that 4 of the 7 largest fires ever were triggered by lightning.

Effect of Climate Change

The progressively increasing severity of wildfires cannot really be in doubt. The data, such as in the figure above, speak for themselves. And, yes, the globe is warming incrementally. Also not seriously questioned is the observation that instances of severe weather have increased, ranging from droughts to floods, as have phenomena such as El Nino and La Nina, which are correlated with severity of weather. The western United States is a reasonable proxy for the relationship between climate change and wildfire severity. The states are arid and depend upon winter precipitation, primarily snowfall, for year-round water for consumption. An increasingly warming climate predicted reduced snowfalls, and earlier snowmelts. This latter statistic was shown to be strongly correlated with areas burned by wildfires. Other observations, such as climate change mediated drought conditions leading to more flammable matter (fuel), have added to the body of belief that climate change was increasing severity of wildfires. Yet, anything approaching causality had been elusive.

Then came the PNAS paper mentioned above.

Courtesy PNAS, Excerpted from Turco M. et al. (2023)

Figure 2. Time series of summer wildfire burn area (BA) and spring through summer monthly average maximum near-surface air temperature (TSmax) for the period 1971- 2021.

The authors used data from California to model area burned in wildfires (BA) against various parameters. Plotted in Figure 2 is burned area (on a log scale) against TSmax, which is the monthly mean of daily maximum near-surface air temperatures. The temperatures were in the period April – October, while the BA was in a smaller summer months subset of May – September. The temperature data are the open circles and black lines, while the BA are the filled circles and red lines.

 The most striking finding is that the single parameter of surface temperature correlated extremely well with area burned*. If there is one parameter that is indelibly linked to climate change, it is global warming, characterized by rise in ambient temperatures. The results of the model are a correlation of 0.84 between the two parameters, with a P-value (a parameter indicating statistical significance) <0.01. Statisticians will find that to be strong. Most of the rest of us can see that it passes the eyeball test. Without exception, the years in which the black circles rise or drop, so do the red ones.  Not all equally strikingly, but follow they do. The data from the first and last three years of the study are extraordinary in this regard. The authors also found that normalizing for precipitation did not make much difference.

            In the end, there is only one measure that will actively address severity of wildfires: slowing down the inexorable march of ever higher near-surface temperatures. Much is happening in that space. More is needed.

Vikram Rao

August 7, 2023

* Everybody look, what’s going down, in For What it’s Worth, by Buffalo Springfield, 1966, written by Steven Stills

References

Turco M. et al. (2023) Anthropogenic climate change impacts exacerbate summer forest fires in California. Proceedings of the National Academy of Sciences 120, e2213815120 https://doi.org/10.1073/pnas.2213815120

Syphard AD, Keeley JE (2015) Location, timing and extent of wildfire vary by cause of ignition. International Journal of Wildland Fire 24, 37–47. doi:10.1071/WF14024