The Texas Forest Service released this graphic showing that more acres have burned this year compared to the same period in 2015. The brown (or orange) bars and line are acres burned in 2016, while blue represents 2015.
The University of Wyoming has issued a publication about the patterns, influences, and effects of wildland fire in the state.
The University of Wyoming paper covers basic facts about fire, weather, intensity, severity, prescribed burning, as well as fire effects and interactions with soils, plants, livestock, wildlife, and bark beetle outbreaks. The document is 16 pages long with an additional 8 pages of references and a glossary. It was written by Derek Scasta, Assistant Professor and Extension Rangeland Specialist.
A couple of items attracted my attention. One is the graphic at the top of this article, the mean fire return interval for Wyoming. If you’re familiar with the geography of an area, data like this can absorb your interest for a while. The map appears to be a section taken out of the whole country map.
Another topic covered in the publication is the relationship between precipitation and acres burned.
The chart above from the paper uses the total Wyoming statewide annual precipitation compared with the total number of acres burned in wildfires each year. We have been thinking that the weather in the summer has a greater effect on acres burned than weather throughout the year. Those weather factors include temperature, relative humidity, wind, and precipitation, and a few others used by the National Fire Danger Rating System. It’s beyond our capacity to analyze all of those, unless we use an index that takes multiple parameters into account, such as the Burning Index or the Energy Release Component.
But what we did (immediately above) was to take one weather parameter from the summer and plotted it on a chart similar to the UofW chart– average monthly precipitation each year for June, July, and August. The weather data came from NOAA, and the acres burned was extracted from the University of Wyoming chart.
Included among the disclaimers is that average precipitation across the state does not apply to every square mile. Thunderstorms in the summer could be hammering one area, while a major fire is burning somewhere else. And, using only precipitation does not take into account temperature, relative humidity, and wind, which are all very important.
If anyone is interested in analyzing the Wyoming fire occurrence data using another weather factor or NFDRS Index (from the summer months), below are the numbers I used. Or, if you’d like to look at another state or geographic area, that would be fine. It’s important to analyze the acres burned and the weather observations for a large area in order get a sample of sufficient size to make it statistically significant. For example, use 15 to 20 years of information from a large national forest with multiple weather stations to reduce the data-skewing impact of a gully-washer thunderstorm at one location.
Far more acres burn in March and April in Kansas than the other months.
Wildfire occurrence data collected by the Kansas Fire Incident Reporting System shows that over an 11-year period, from 2004 through 2014, there were two spikes in the number of reported fires during the year — in March/April and in July. However, by far, most of the acres burned in March and April.
Below are excerpts from a Kansas State University article written by Chip Redmond and Mary Knapp of the University’s Weather Data library.
“…Of the eleven years of recorded KFIRS data, 2014 had both the most land area burned (187,500 acres) and the highest number of reported fires (8,075) in a year. In both land area and number, fires followed the typical curve of the previous 10 years, with a peak during the spring months of March and April.
This trend of spring wildfires coincides very well with prescribed burning season. Early spring is typically characterized by dead/cured fuels, warming temperatures, low relative humidity, and little precipitation. Often, if the late fall and winter are below normal in precipitation and/or the region is in a drought, these spring conditions are enhanced – providing explosive conditions for fire growth.
This was the case entering 2014. However, conditions were worsened by an increased fuel load from late summer rains of 2013 that brought some relief to the prolonged drought which peaked in 2012.These combined factors led to a peak of 156,600 acres burning in March/April 2014 alone. Of the 61 days possible in March and April, almost half (29) had more than 1000 acres burned statewide each day. Two days (Jan. 26 and Feb. 19) occurred outside of these months. Below, the March/April calendar with red days were those in which more than 1000 acres statewide were burned. Cold frontal passages throughout the period are marked on their associated day. Continue reading “March and April are historically busy for wildland firefighters in Kansas”
For the last several days we have been writing about fatalities on wildland fires — the annual numbers and trends going back to 1910 and some thoughts about how to reduce the number of entrapments (also known as burnovers). Often when we think about these accidents, what automatically comes to mind are the entrapments. When multiple firefighters are killed at the same time it can be etched into our memory banks to a greater extent than when one person is killed in a vehicle rollover or is hit by a falling tree. Much of the nation mourned when 19 members of the Granite Mountain Hotshots were overrun and killed by the Yarnell Hill Fire in Arizona in 2013. A fatal heart attack on a fire does not receive nearly as much attention.
When we discuss ways to decrease deaths on fires, for some of us our first thoughts are how to prevent entrapments, myself included. One reason is that it can seem they are preventable. Someone made a decision to be in a certain location at a specific time, and it’s easy to think that if only a different decision had been made those people would still be alive. Of course it is not that simple. Perfect 20/20 hindsight is tempting for the Monday Morning Incident Commander. Who knows — if they had been there with access to the same information they may have made the same series of decisions.
An analysis of the data provided by NIFC for the 440 fatalities from 1990 through 2014 shows that entrapments are the fourth leading cause of fatalities. The top four categories which account for 88 percent are, in decreasing order, medical issues, aircraft accidents, vehicle accidents, and entrapments. The numbers for those four are remarkably similar, ranging from 23 to 21 percent of the total. Number five is hazardous trees at 4 percent followed by the Work Capacity Test, heat illness, and electrocution, all at around 1 percent. A bunch of miscellaneous causes adds up to 4 percent.
NIFC’s data used to separate air tanker crashes from accidents involving other types of aircraft such as lead planes and helicopters. But in recent years they began lumping them all into an “aircraft accident” category, so it is no longer possible to study them separately. This is unfortunate, since the missions are completely different and involve very dissimilar personnel, conflating firefighters who are passengers in the same category as air tankers having one- to seven-person crews — from Single Engine Air Tankers to military MAFFS air tankers.
The bottom line, at least for this quick look at the numbers, is that in addition to trying to mitigate the number of entrapments, we should be spending at least as much time and effort to reduce the numbers of wildland firefighters who die from medical issues and accidents in vehicles and aircraft.
Last week Secretary of Agriculture Tom Vilsack reported that 13 wildland firefighters lost their lives in the line of duty in 2015. That was an increase from 2014 when there were 10 fatalities, and was about a third of the 34 that were killed in 2013 — that year included the deaths of 19 members of the Granite Mountain Hotshots near Yarnell, Arizona.
The National Interagency Fire Center has statistics about line of duty deaths going back to 1910. During that time, according to their numbers, 1,099 firefighters died.
In looking at the 105 years of NIFC data there appears to be an increasing trend. The figures below are the average number of fatalities each year for the indicated time periods:
One likely explanation for the apparent increase is that 80 to 105 years ago probably not all fatalities were reported or ended up in a centralized data base, especially those that occurred on state or locally protected lands. Even if we only look at the figures since 1960, as in the chart above, it still shows a steep increase over those 55 years.
It is possible in the last 25 years the reporting of fatalities and the collection of the data has been somewhat more consistent and complete. The chart below covers that period, from 1990 through 2015, and has a slight downward trend, which would be even more obvious if not for the 19-person crew that passed away in 2013 on the Yarnell Hill Fire.
I can’t prove that there was under-reporting of wildland firefighter fatalities during most of the 20th century, but if a firefighter was killed on a vegetation fire in Missouri in 1921, I can see how that statistic may not have made it into the data base that is now maintained at NIFC.
So what does all this mean? Individuals can look at the same batch of statistics and develop vastly different interpretations. However, it would not be prudent to assume that the fatality rate almost tripled from the first part of the 105-year period to the last 25 years. There are several ways to analyze data like this. The least complex is to look at the trend of the raw numbers of fatalities year to year. A more complex and meaningful method would be to determine the fatality RATE. For example, the fatalities per million hours spent traveling to and working on fires. That would be impossible to ferret out during most of the last 105 years. But the firefighting agencies should be able to find a way to begin collecting this information, if they don’t have it already.
If the fatality and serious injury rates were calculated over a multi-year period, it should illustrate the effectiveness of a risk management program. Otherwise, the simple number of deaths each year might be affected to an unknown degree by the number of acres burned. Other factors could also affect the numbers, such as fire intensity influenced by fuel treatment programs, fire history, drought, climate change, or arson.
Should firefighting agencies have specific goals about serious injuries and fatalities? Is there an acceptable number? Is 5 a year too many? Is 15 too many? Is it stupid to have a goal of zero fatalities — or any number?
The chart below superimposes the number of fatalities over the acres burned in the United States from 1990 through 2015, but it does not include Alaska since many fires there are not suppressed, or they are only suppressed in areas where they threaten structures or people. In 2015 more acres burned in Alaska than all of the other states combined.
UPDATED January 17, 2016
One of our loyal readers, Bean, has been thinking about this issue and figured that since the amount of firefighters’ exposure to risk is necessary in order to calculate trends, perhaps parameters other than acres burned could be correlated with the number of fatalities. Data that is publicly available as far back as 1990 or 1994 includes mobilizations of incident management teams, crews, overhead, helicopters, air tankers, air attack ships, infrared aircraft, MAFFS air tankers, caterers, military firefighters, and shower units. I considered all of those and concluded that the number of crews mobilized would come the closest to serving as a proxy for accurate data of how many hours all firefighters spent traveling to and working on fires.
Data for crew mobilizations is available from 1990 through 2014. I divided the number of crews mobilized by the number of fatalities for each year and called this the Fatality/Crews Mobilized Index.
Like the earlier chart comparing fatalities to acres burned, this analysis also shows a decreasing trend in the last 25 years. In a comment posted January 17, Kevin9 said the earlier acres/fatalities analysis is “spiky”. This newer crews mobilized/fatalities data also has spikes (especially in 1997 and 2009) but not quite to the degree the earlier chart had. During the 25-year period, 1997 had the least number of acres burned and crews mobilized, but still had 10 fatalities. The second lowest number of crews mobilized occurred in 2009 and there were 15 fatalities that year.
As an experiment, knowing that there were mass casualty events in 1994 and 2013 (14 and 19 fatalities respectively), just to see what the effects were, I changed the data in those two years to the average for the last 25 years, which is 17, and there was no major change in the trend line, except it was a little lower across the entire range.
It’s been a long time since I took statistics courses, but here’s what I came up with when analyzing the Fatality/Crews Mobilized Index data:
- Standard deviation: 0.019
- Mean: 0.026
- Coefficient of variation: 0.770
- Variance: 0.00037