Bad Year for Wild Fire

We’ve already seen an over-abundance of area burned by wildland fires in the U.S. As bad as it’s been in the western states, it’s been even worse in Alaska. And Canada has, unfortunately, had to write a new record book.

Here’s the area burned (in millions of acres) for each year from 1960 through 2014, which does not include this year (data from NIFC, the National Interagency Fire Center):


It shows a significant upward trend over time; wildfire in the U.S. is getting worse.

But it doesn’t show this year, because the year isn’t over yet. NIFC does, however, provide data for year-to-date area burned, but only since 2005. Let’s add that to the graph:


It’s been pretty high this year, in fact we’ve already suffered more area burned that the whole-year average from 1960 through 2014.

If we compare the area burned before July 17 to that burned after, there are hints but no obvious relationship:


Linear fits (both parametric using least squares, and non-parametric using Theil-Sen) again show hints, but without a statistically significant relationship. I guess that’s to be expected when one only has data since 2005.

Any way you look at it, we can expect more wildfire this year. Not only California, but much of the west is under the thumb of drought. Probably worst is the fact that the snowpack in the west is tremendously low. Snowpack is a crucial source of water in western states, but it’s simply not there this year.

We can certainly expect this year to be a big wildfire season, possibly a record-setter (but prediction is hard, especially about the future). It will be (already has been) a severe strain on firefighters and resources, and a terrible drain on our economy because fighting so much wildfire costs money. And, even with the best — our firefighters — you can’t stop all wildfires from destroying homes and businesses, and taking lives. What’s the price tag on that?

15 responses to “Bad Year for Wild Fire

  1. Dick Veldkamp

    Is there an explanation for the fact that ‘acres burned’ goes down until 1985? I could imagine there is influence of better observation methods, improved fire control, or something like that?

    Or is this just a coincidence (i.e. not statistically significant) ?

    [Response: Not statistically significant.]

    • Reading Tamino’s post I was going to make a point that is similar to the one that Dick raised, which is that the increase in wildfire burning occurs in the face of increased effectiveness of fire-fighting technology. One could probabaly assume that all other things being equal, the increase would actually be steeper without the impact of improved technologies/methodologies.

      At some point though in the warming of the planet the capacity of response to wildfire is likely to be overwhelmed.

  2. Horatio Algeranon

    “Emissions decreased with the global economic recession: 77% of the decrease 2007-2009 was due to decreased consumption and changes in the production structure of the US economy, with just 18% related to changes in the fuel mix of the energy sector…. changes in fuel mix—the primary means by which substitution of gas for coal affects emissions—have had a relatively minor role in the reduction of US CO2 emissions since 2007”

    Drivers of U.S. CO2 emissions 1997-2013. Available from:

  3. Besides tamino’s observation re. significance, these data do come with the following disclaimer:

    The National Interagency Coordination Center at NIFC compiles annual wildland fire statistics for federal and state agencies. This information is provided through Situation Reports, which have been in use for several decades. Prior to 1983, sources of these figures are not known, or cannot be confirmed, and were not derived from the current situation reporting process. As a result the figures above prior to 1983 shouldn’t be compared to later data. (

    It is possible this disclaimer should be published more saliently by the USFS. That said there is no evidence that there is a serious bias here.

    • Hmm. That raises the possibility of a reporting bias, though–from other reading, fire policy shifted gradually to allow fires to burn more often. If fires were deemed non-threatening, and were monitored only, would there be incident reports? Dunno, myself.

  4. Thanks, Tamino. I’d been wondering if we’d see a wildfire uptick, even dropping by the NFIC site a couple of times.

    I guess it’s appropriate that the early trend Dick comments on turns out not to be significant; there’s a mix of contradictory ‘forcings’, including climate, technological change in firefighting, and fire suppression policy (which changed markedly between, say, 1965 and 1990.)

  5. thefordprefect

    shouldn’t this be % of total fire risk area (excludes deserts, recently burnt etc) burning. Eventually with all scrub burnt the fire risk should be zero! Or is this just insignificant?

  6. Some docs that are worth a look.

    “Earth on Fire” (ABC, Catalyst production)

    “Forests under Fire” (Shows an interesting forest restoration project, reducing tree density to pre-settlement conditions (partly for fire management). Cut down smaller trees, some goes to sawmill, some used as biomass. One of those projects that could be really good or really bad, depending on how it is carried out/who is in charge… but horrible production values, made even worse if you like to watch docks at higher speeds)

  7. Tamino, I learn so much from this web site. Thank you. I am trying to understand the difference between type 1 and type 2 errors… and mostly the ramifications of misunderstanding.

    It came up early in the new risk document : Climate Change: A Risk Assessment
    See Sir David King’s Video on the Climate Risk Report -
    Climate change risk hashtag #climateriskreport.

    “Scientists are conservative about drawing incorrect conclusions—so much so that they would rather draw no conclusion than an incorrect one. Consequently, they have developed standard practices and cultural
    norms to protect the scientific knowledge pool from being contaminated by falsehoods. For example, scientists typically apply statistical tests that estimate the probability that a predicted outcome may have happened purely by chance rather than because of a hypothesized cause. If the probability of the random outcome is greater than five percent, standard practice is to reject the hypothesis. Ironically, this rigor often results in the rejection of a correct hypothesis because there was only a small chance—potentially less than 6 percent—that the hypothesis was indeed a random outcome.

    Such scenarios involve two types of uncertainty, or ‘error’ in statistical terminology. First is the possibility that the hypothesized cause is accepted, but is actually wrong. This condition is commonly called a‘false-positive;’ statisticians call it a ‘type I error.’ Conversely, there is the possibility that the hypothesis is rejected, but is actually correct. This situation presents a false-negative, or ‘type II error.’

    Scientists. See,are relatively tolerant of false-negatives: in most scientific fields it is not standard practice to estimate the probability of committing a type II error.In contrast, professional risk managers are often more concerned about type II errors which could result in their disregarding a risk with potentially severe consequences. For example, even though the
    probability of any particular house burning down in a given year is very low, the mortgage lender requires the homeowner to carry casualty insurance to protect the lender’s investment. The point is that even a very low probability of an outcome may represent a large risk if the outcome would be very severe.

    Consequently, when scientists tolerate type II errors, their work may lack rigour from the standpoint of the decision-makers they seek to inform.

    Downward bias under uncertainty Consistent with their aversion to type I error and tolerance of type II error, climate scientists have often erred toward underestimating risk when faced with deep uncertainty. A stark illustration of this phenomenon occurred when the IPCC’s ‘Reasons for Concern’ (RFC), first published in 2001, were updated in 2009. The RFCs are categories of climate change impacts that IPCC authors deemed of
    potential interest to decision-makers and include risks to unique and threatened ecosystems, extreme weather events, distribution of impacts geographically and across income classes, aggregate economic impacts, and sudden dramatic changes in the regulation of the global climate (e.g., a sudden collapse of a large ice sheet leading to abrupt sea level rise). The RFC assessment evaluated the sensitivity of each RFC category to global temperature increases between 0 and 5 degrees Celsius.”

    large ice sheet leading to abrupt sea level rise). The RFC assessment evaluated the sensitivity of each RFC
    category to global temperature increases between 0 and 5 degrees Celsius.

  8. I posted this picture over at hotwhopper a couple weeks ago when it was really bad. On a clear day, you can see the Alaska Range about 150 miles away. I think the visibility on that day was less than 1/4 mile.

  9. Dan Andrews

    Quite dry in BC too. The rain forest itself is dry. Some of the islands are under water advisories with well levels now at the level they typically are in September. Lower mainland reservoirs are in “uncharted territory” according to advisors, never having been so low. The day I arrived here you could look directly at the noonday sun in Vancouver as the smoke from northern fires was so thick. Lawns are dry and brown except for those who water them despite restrictions (we know who you are…emerald green lawns stand out amidst brown ones). In three weeks we’ve had one bit of rain overnight, hardly enough to make a difference. Usually this area is a lush verdant green. Definitely an unusual year. One the other hand the lawn doesn’t need mowing.

    • A few weeks back in South Carolina we observed a lovely ‘pink’ full moon–to my eye, closer to copper-colored. Reportedly, it was due to smoke from Canadian wildfires. I thought BC at the time, but it was also around the time that Saskatchewan had such an outbreak, I think.