Issue 7

Time to Burn: Getting a Step Ahead of Wildland Arsonists

by Zoë Hoyle

Every year, arsonists set over 1.5 million fires in the United States, resulting in over $3 billion in damages. Arson is a leading cause of wildfire in several heavily populated States—Florida, for one. Often set near homes and roads, intentionally set fires cause a disproportionate amount of the damage attributed to wildfire in general.

Research forester Jeff Prestemon from the SRS Economics of Forest Protection and Management unit in Research Triangle Park, NC, and economist David Butry, formerly with the SRS unit and now with the National Institute of Standards and Technology in Gaithersburg, MD, have developed a model to help law enforcement agencies better predict where and when fires might be set in wildland areas—and design strategies to reduce the risk of arson.

Criminal Dimensions

For other types of crimes, researchers have documented that perpetrators often commit multiple offenses in a short time frame, a “spree” phenomenon described as temporal clustering.

Individuals committing property or violent crimes also often commit multiple crimes within a certain area, adding a second dimension of spatial clustering. In a pair of studies, Prestemon and Butry set out to test whether spatiotemporal clustering could also be observed in wildland arson. They also looked at the relationship between socioeconomic factors and incidences of wildland arson—adding yet another dimension by describing the firesetting process in the context of the economics of crime.

“Even though the economic damages from wildland arson are often staggering, research into the factors that contribute to it has been limited to a few published studies,” says Prestemon. “Models of wildland arson have mostly related fire setting to weather, seasonal trends, and law enforcement, ignoring the socioeconomic variables used to predict other types of crime.”

“At the same time, no one had previously identified the spatiotemporal dimensions of wildland arson that we found,” says Butry. “Our findings have uncovered a new avenue of fire research, deepened our understanding of arsonist behaviors, and revealing another way in which humans and society interact with the environment.”

Patterns Emerge

Two studies revealed similar spatio-temporal patterns for arson as for other crimes. In one study in Florida, Butry and Prestemon evaluated wildland arson as both an annual and a daily process. Using annual data from all Florida counties for 1995 to 2001, their model revealed the influence of law enforcement, wildland fuels, poverty, and labor conditions on the rates of ignitions recorded for the State.

To measure fine temporal patterns of arson, the researchers parsed the 1995 to 2001 data into daily observations. They focused this finescale analysis on nine counties in Florida with high rates of arson. Their model identified temporal clustering that lasted up to 11 days—implying higher risk of repeat arson ignitions for 11 days following the initial fire. This kind of pattern had never before been found in any research into humanignited wildfires.

In a second study, Butry and Prestemon measured the spatial as well as temporal clustering of arson wildfires using a different set of data—daily information for the six U.S. census tracts in Florida with the greatest arson activity. Their statistical results showed that an arson event in one census tract was related to arson in the neighboring tracts for up to 11 days, and in the same tract for up to 10 days.

“Not only did we confirm our findings from the previous study, but we also showed that arson clusters in both space and time,” says Butry. “In other words, you can use arson events in one tract to predict future ignitions in the same or adjacent tracts for several days.”

How to Get Ahead of Wildland Arsonists

“Combining the patterns we found with data on law enforcement led us to conclude that there are strategies that law enforcement can use to prevent wildland arson,” says Prestemon.

“They can closely monitor areas where fires have been set before. They can also increase arson enforcement on days of the year when events are more common, and during droughts.”

From their studies, Prestemon and Butry also found that locations with difficult economic conditions—low wages and high poverty rates—have higher rates of wildland arson, a finding consistent with other economic models of crime. They also found that forest management activities are related to wildland arson, with fuel reductions from prescribed burning and other wildfires correlated with lower arson rates.

“This finding is also consistent with an economic model of wildland arson crime, where lower fuels increase the cost of successfully starting fires,” says Prestemon.

The next step is to test the model in other locations to see if the statistical results hold true in other States, and even further, to conduct new research on what motivates people to set damaging fires. “We need to direct some of our effort into understanding who arsonists are and what, precisely, makes them behave the way they do,” says Butry. “If we could add feedback from convicted wildland arsonists, we could really enhance our understanding of how they choose where and when to set fires. Incorporating this information into the model would raise its predictive value to law enforcement and give more insight into an important social problem.”

For more information

Jeff Prestemon at 919-549-4033 or jprestemon@fs.fed.us

David Butry at 301-975-6136 or David.Butry@nist.gov