Are Wildfire Arson Cases on the Rise?
By Scott Fischer, SWAIT Contributor
House Fresh data journalists G. John Cole and Danny Ashton examined fire “cause” data for 2022 and 2023 in their article “The U.S. wildfire index: How and where wildfires spread across America.” As stated in their article, they used data from the National Interagency Fire Center (NIFC). The NIFC pulls data from an application called InFORM. InFORM is used by federal fire agencies and some state and local fire agencies to collect statistics on wildland fires. While most wildland fires occurring on federal land management jurisdictions should be in the data set for a given year, the data does not represent all wildland fires occurring in the United States.
The House Fresh investigation broke down fire numbers by cause comparing data from 2022 and 2023. While most of the fires over the two-year period were unintentional human caused fires, they found there to be 658 fires categorized as arson and 428 categorized as undetermined out of 6707 reported fires. This represents approximately 9.8% of fires being arson and 6.4% undetermined for the two-year dataset.
When looking at the numbers by year, there were 3340 reported fires in 2022 and 3367 fires in 2023. This represents only a small increase in fires (<1%) over the two years. Yet when looking at arson fires, they found 294 in 2022 vs 364 in 2023 within the dataset. When comparing just the total number of arson fires, there is a 23.8% (334-294/294) difference between 2022 and 2023. Undetermined fires remained nearly the same over the two years with 211 in 2022 and 217 in 2023.
Have we really seen a nearly 24% increase in arson fires? Have our investigators gotten better at identifying arson fires, or have our investigators gotten better at classifying fires as arson over the past two years? The answer is likely to be a combination of all three.
Many new and inexperienced fire investigators struggle with understanding the difference between being able to determine the ignition source of a fire versus determining fire cause. Incendiary (arson) devices, while more common for wildland arsonist vs. urban arsonists, are still rare to find in the wildland. Most wildland arson fires are “hot set,” leaving no physical evidence of the ignition source. Inexperienced fire investigators may classify a fire as undetermined simply because they were unable to determine an ignition source while there was other evidence of arson present at the fire scene or based on witness information.
The National Wildfire Coordinating Groups PMS412, Guide to Wildland Fire Origin and Cause Determination, Chapter 7:Arson Factors, describes how a fire investigator can use other evidence to help determine a fire cause of arson while unable to determine an exact ignition source. All Wildland Fire Investigators trained under the NWCGs FI210 course are expected to use PMS412. PMS412 is currently being updated, including Chapter 7: Arson Factors. The update is expected to be published by spring of 2025. As fire investigators gain more experience and become confident in their skills, we generally will see investigators classifying fewer fires as undetermined.
When it comes to classification of fires, the NWCG has created some confusion over the past few years. Back in 2020, NWCG revised the wildfire cause data standard used for reporting fire cause. The fire causes are broken down into seven different data fields. The primary four data fields are Cause Classification, General Cause, Specific Cause and Specific Cause Detail. While Cause Classification is either natural, undetermined, or human, the other data fields are rich with a cascade of different fire causes.
The confusion in cause classification can happen in several ways. Currently, the NWCGs wildland fire investigator course, FI210, teaches the old cause classification. The course has not been updated since 2016. NWCG is currently working on an update to FI210 alongside the PMS412 update. The updated versions are expected to be released in early 2025. These updates will address the new NWCG wildfire cause data standard.
Additionally, most fire investigators do not enter their own stats into the system of record for fire cause. NIFC, the agency responsible for the data used in the House Fresh article, pulls data from InFORM. Most of the statistics related to fire cause are reported in the system by non-investigators or supervisors/managers. The people entering the information may not have the experience to translate what the investigator put in their report as the basic fire cause to match the new NWCG fire cause classification. Additionally, in 2022 and 2023, InFORM required all fires be closed out and locked within 30 days of the fire. Many fire investigations may go longer than 30 days. Managers simply put “undetermined” into the system to close out the fire. This resulted in a higher number of undetermined fires being reported in the system.
To fix these issues, in 2024, NWCG and the InFORM admins worked to add “under investigation” to the Cause Classification category into the system. Managers who wish to close out their fires in InFORM can then select “under investigation” as opposed to “undetermined”. InFORM will then remind the person responsible for entering “cause” category to change the fire cause even if the fire has been closed out.
What about the undetermined fires? How many of those are arson? There are many reasons a fire can be classified as undetermined. A fire investigation is based on the scientific method of inquiry. Investigators gather available evidence and information to determine potential hypotheses for fire cause. If an investigator is unable to narrow down their hypotheses based on the evidence or information determined during the investigation, to a single hypothesis, they would classify the fire as undetermined. The reality is these undetermined fires were caused by something and would fit within one of the cause categories if there was evidence discovered to classify the fire. It is likely the undetermined fires, if known, would be spread amongst all the cause categories based on percentage of fires per category within the original dataset. Based on the 2022 and 2023 fire data, arson may go up as much as 42 fires (~10% of 428 fires) bringing the total to 700 arson fires, or an increase of less than 1% (10.4 vs 9.8%).
Not all fires are investigated. In theory, regardless of whether a fire was investigated, all fires are entered into InFORM by agencies that use the application. InFORM pulls all its initial fire statistics from the IRWIN database. IRWIN is linked to the federal fire dispatch centers. Any fire a federal fire agency responds to should therefore be in InFORM. Since the system doesn’t allow for reporting fires as “uninvestigated”, they are classified as undetermined. This adds to the undetermined dilemma of fire classification within InFORM.
What about all the other categories? How many of those fires are actually arson? Wildland fire investigations are complex and not all fire causes are correct. This is a hard pill to swallow within the industry. Looking at serial arson case files, there are numerous examples of fires being classified as something other than arson to be later reclassified as arson after further investigation, sometimes years later. Fire investigation requires highly skilled investigators. As stated above, new, and inexperienced investigators may not get it right. The hope is they err to the side of undetermined but that is not always the case.
Simply looking at the statistics reveals a significant increase in arson reported fires over the two-year period from 2022 through 2023. But statistics can be deceiving. Based on the above information, there was an increase in arson fires but most likely not as significant as the raw statistics will lead you to believe. As an industry, we need to do a better job and invest more time in understanding how to categorize and enter data correctly into the fire cause dataset.