Why Virtual Fire Scenes?
Updated: Sep 9
A Better Way to Train Fire Investigators
One of the biggest challenges in training fire investigators is pretty simple: good fire scenes are hard to come by.
We can teach fire dynamics, fire patterns, ventilation, and origin determination in a classroom. We can show photographs and diagrams. But eventually, an investigator needs to put all of that information together and work through a fire scene.
Ideally, we would train investigators by giving them an unlimited supply of realistically furnished houses, burning them under controlled conditions, and then letting them investigate the resulting scenes. Obviously, that isn't very practical.
So how do we give investigators experience with a large number of realistic fire scenes without having to burn a large number of houses?
That's where Origin 360 comes in.
Origin 360 Brings the Fire Scene to the Investigator. The virtual fire scenes on Origin 360 use 360-degree photography, 3-D modeling, and other scene documentation techniques to allow investigators to explore an actual fire scene remotely.
Instead of looking at a handful of photographs in a PowerPoint presentation, the investigator can move through the structure, look around rooms, examine patterns from different viewpoints, and understand how different pieces of evidence relate to one another spatially.
That last part is important...
A photograph can show you a fire pattern.
A virtual scene can show you where that pattern is in relation to the ventilation openings, fuel packages, other patterns, and the rest of the structure.
My graduate research at Oklahoma State University examined this concept as a potential solution to some of the limitations of conventional fire investigation training and research. Virtual scenes allow investigators to examine an entire fire scene remotely and evaluate spatial relationships that can be difficult to understand from still photographs alone.
One Fire Can Train Hundreds of Investigators
This is where virtual scenes become particularly resource efficient.
A live burn may require a structure or burn building, furnishings, instructors, firefighters, safety personnel, equipment, fuel, setup time, and cleanup. At the end of the exercise, the scene eventually disappears. A virtual fire scene doesn't.
Once a fire has been properly documented, that scene can become a reusable training resource.
Ten investigators can work through it.
A hundred investigators can work through it.
Investigators in different states can work through exactly the same exercise without traveling anywhere.
And the scene can still be used years later.
This doesn't eliminate the need for hands-on training. Investigators still need experience excavating scenes, examining physical evidence, taking measurements, and performing the other hands-on tasks associated with an actual investigation.
The goal isn't to replace live-fire training. The goal is to make it possible for investigators to experience far more fire scenes than live-fire training alone could ever practically provide.
Experience Matters—So Let's Give Investigators More of It
Think about how investigators gain experience in the real world.
We investigate fires as they occur. That means the investigator has little control over how many scenes or the kinds of scenes they encounter. An investigator might only see a few relatively straightforward kitchen fires in a year but very few heavily damaged, ventilation-controlled fires where the origin is difficult to determine.
With a library of virtual scenes, we can intentionally expose investigators to different problems.
One exercise might involve a relatively straightforward pre-flashover fire.
The next might involve multiple rooms. Another might contain significant post-flashover damage and competing fire patterns. Another might involve an elevated origin (Z-factor) fire. Another might contain a ventilation-generated pattern that looks very convincing until the investigator considers the rest of the scene. Instead of waiting years for investigators to encounter those situations in the field, we can bring those situations to them.
But There Is Another Reason I'm Interested in Virtual Fire Scenes Research.
Fire investigation is a forensic discipline, and the methods we use to determine fire origin should be testable.
One surprisingly difficult question is...
If we give a large number of trained fire investigators exactly the same fire scene, how often will they reach the same conclusion?
There isn't as much research answering that question as you might expect.
One of the largest published origin-determination studies involved 586 professional investigators examining photographs from the same test fire. That study provided important information about investigator performance, but it was still based on a single fire scenario.
That leaves a lot of unanswered questions.
What happens as fire damage increases?
What happens when ventilation becomes more complicated?
Do investigators agree more often on some types of scenes than others?
Does experience affect performance?
Does certification?
Do certain analytical methods produce more consistent conclusions?
Answering those questions with live burns would require conducting numerous fires and getting a statistically meaningful number of investigators to examine each one.
That's enormously difficult and expensive.
Virtual fire scenes change the equation.
The Fire Scene Becomes a Research Tool
Imagine 200 investigators working through the exact same Origin 360 exercise.
Each investigator examines the same structure and the same fire patterns. They can then identify their area of origin and explain how they reached that conclusion.
Now we have data.
Maybe most investigators identify roughly the same area.
Maybe the responses are scattered throughout the room.
Maybe investigators consistently divide into two groups.
And that's where things get interesting.
Why did they disagree?
Perhaps a ventilation-generated pattern pulled investigators toward a doorway.
Perhaps one particular fire effect was given too much weight.
Perhaps investigators using an origin matrix reached different conclusions than investigators who didn't.
Virtual scenes make it possible to study those questions because large numbers of investigators can examine essentially the same evidence without ever having to travel to the fire scene. That was one of the central ideas behind my graduate research.
Wrong Answers Can Be Valuable
This may be one of the most useful aspects of the whole approach.
Suppose 200 investigators complete an exercise and 150 identify the correct origin.
It would be easy to focus on the 150.
I'm also interested in the other 50.
What did they see that led them somewhere else?
If investigators repeatedly make the same mistake on the same type of fire scene, we haven't just identified 50 wrong answers.
We've identified a potential weakness in our training or methodology.
And now we can study it.
We Can Examine More Than Origin Determination
Once we have a library of virtual fire scenes, the research possibilities extend well beyond simply asking investigators, "Where did this fire start?"
The same scenes could potentially be used to evaluate investigative tools such as heat and flame vector analysis, origin matrices, and depth-of-calcination measurements.
They could also be used to study cognitive bias. For example, two groups of investigators could receive the exact same fire scene but different contextual information. If their conclusions differ, we can begin examining how that information affected their interpretation of the physical evidence. We could then test strategies intended to reduce those effects, such as contextual shielding or sequential linear unmasking.
A Fire Scene That Doesn't Disappear
Real fire scenes are temporary. Eventually the debris is removed. The structure is repaired or demolished. The evidence changes, and the scene disappears. A properly documented virtual scene can be preserved. That means a fire that happened today could still be training investigators five or ten years from now.
More importantly, investigators in different places and at different points in their careers can examine the same scene. That gives us something that has historically been very difficult to achieve in fire investigation: a practical way to repeatedly expose large numbers of investigators to standardized, realistic exercises.
That's the larger idea behind Origin 360.
It isn't intended to replace the fire scene. It isn't intended to replace live burns.
And it certainly isn't intended to replace hands-on investigation. It's a way to give investigators access to more scenes, more difficult problems, and more opportunities to practice making and defending origin determinations.
At the same time, those exercises can help us collect something the fire investigation profession badly needs: data about how investigators actually interpret fire scenes. Instead of trying to bring every investigator to the fire scene, Origin 360 brings the fire scene to the investigator.
Comments