Q&A: Expert sheds light on difficulty of finding children missing in the woods

After a six-day search for two missing children wrapped up this week in Pictou County, an expert on lost people is shedding light on the difficulty of finding children in the woods.

Lily Sullivan, 6, and Jack Sullivan, 4, have been missing since May 2, when police received a 911 call reporting that they had wandered away from their home on Gairloch Road in Lansdowne Station, roughly 25 kilometres southwest of New Glasgow, N.S.

Police announced Wednesday that after six days of hundreds of volunteers scouring the heavily wooded areas surrounding the home, there was no sign of the children and it's not likely they are alive.

Robert Koester, a search mission co-ordinator with the Virginia Department of Emergency Management in the United States, wrote the book Lost Person Behavior: A Search and Rescue Guide on Where to Look - for Land, Air and Water. He also created an app called Lost Person Behavior, which was used by searchers in Lansdowne Station. 

Koester has collected a database of a half-million search and rescue incidents from around the world and compiled them in a database, which he uses to create statistics. 

He told CBC News he can't speculate on the Lansdowne Station search but can provide general comments based on his data. 

This conversation has been edited for length and clarity:

When a four-year-old and a six-year-old child go missing in a rural, wooded area, what are the first things that searchers need to consider?

Well, the first thing is the beginning of a good solid investigation: the age, the scenario, where were they last seen, when they went missing, the timelines, what caused them to go missing? Usually there's not one possible scenario, but multiple possible scenarios that could be considered. And by scenario, I mean they could have gotten lost … they could have gotten hurt, they could have gotten sick, although two children at the same time getting sick would be much less likely. They could have gone into a water feature if water features are a possibility, and a criminal scenario is also always a scenario. All these things have to be considered.

Usually, the next thing is once you know the age group is we can look at some basic statistics like how far do four- to six-year-olds typically go when they get lost … Half of four- to six-year-olds are going to be found within … a half a kilometre away.

So 75 per cent [of children] are 1.2 kilometres away.… If you want to go far enough to find 95 per cent then there's a dramatic jump because you're now starting to get into more outliers and that number is 6.6 kilometres away, which dramatically increases your search area.

And what about the forest conditions? This area has really dense underbrush and fallen trees, and it's almost like you can't even see through the forest when you're trying to look. How could that set kids back in terms of how far they can go?

This sometimes limits the data, but in kind of a surprising result, a lot of my data also comes from the Pacific Northwest, so Oregon, Washington state, British Columbia. And you would think those thicker forests might slow people down, but in fact they tend to go a little bit further than in the more eastern vegetation.

Certainly in Virginia, which is where I'm based out of, people don't go as far during the summer because of all of the thick vines and thick vegetation. So it's always taken into consideration, but the data includes all of those cases.

I saw [your app includes] chances of survivability for kids who are missing. Can you go over that based on the amount of time that they're missing?

Essentially if you find a child in the first 24 hours, their chance of survivability is going to be pretty good. As a matter of fact, 98 per cent are found alive if found in the first 24 hours. Fatalities tend to be drownings. If you don't find them until two days have gone by, it drops down to only 33 per cent are found alive.

You occasionally get some of those miraculous finds. Now, what my chart doesn't take into consideration is temperature and weather. And obviously temperature and weather is going to have a huge influence on the actual survivability.

So this search … they called off the search and rescue portion after six days. When do searches normally call off that portion of the operation?

I imagine they didn't call it off, but they suspended the search. And if there's new information they will certainly be out there searching again. Turns out the average time of a suspended search is just one day. Now that's an average, some searches will go two days, some will go a week. Everything is obviously a little bit different depending upon the circumstances. But if they went a week, they certainly searched a lot longer than average.

And how often is it that there are clues? These kids didn't leave any trace. [RCMP said] they thought they saw one boot print and then there was nothing else. So how common is that?

So most searches will have at least one or two clues … I have certainly run searches with no clues, the only clue was the last place they were seen, and then we find the subject. So it's really hard to say. It also depends upon, you know, do you have trained trackers? We usually call a dog alert a clue.

I've seen both — I find the subject with no clues, and I have lots of clues and never find the subject.

How often do they never find the subject alive or find a body?

Well, never finding the subject, which we would call it a search suspension, that occurs about five per cent of the time in the database.

You mentioned tracking dogs — how often is it that they don't catch any scent?

That's kind of hard to evaluate because to give a real percentage, you have to know both how many times they caught the scent, how many times they didn't catch the scent. And then the more important part of the equation was how many times was there scent out there to catch? You're not going to catch any scent if the subject's not there.

You may also miss the scent if the wind's blowing the wrong direction, the dog has a cold … no resource is absolute.

And you mentioned at the start [of the interview] that sometimes there is criminality involved. What percentage of the time is that the case?

So looking at my database, if it is an urban search … four per cent of the time there's criminality. If it is a wilderness or rural search, it's one-third of one per cent, so relatively rare.

What are the chances that the drone won't pick up any heat or won't find the person if they are there in the woods?

So the two primary sensors that are used is an optical camera, and the second is an infrared sensor. Both sensors have to be able to see the ground. In other words the [tree] canopy will stop them. Easy to understand how canopy would stop a visual sensor, but it turns out infrared heat is also stopped by leaf cover.

The second part is what is the visibility of the subject … for a normal camera, if there's high contrast, pretty good chance of seeing it. If there's low contrast, even with no canopy, can be very difficult to spot. And for infrared that starts to get into heat patterns, you know if they're severely hypothermic and at the same temperature as the environment, then there's little chance to be spotted. Certainly if they're under canopy, they're not going to be spotted either.

And then in wilderness or rural searches, how often are animals involved? Are there ever animal attacks that are the cause of the missing person?

That is rare. Especially with two individuals, that's extremely rare. However, I do see it in the database, but you know, if there was an animal attack, it was usually something like a mountain lion with a single individual who was out there. The typical scenario for an animal attack was early in the morning or late in the afternoon with a runner running by themselves on a wilderness trail, and running kind of triggers that attack phase to an animal.

When a person that's being searched for is never found, what are usually the causes? Or can you ever know the causes of why they weren't found?

I have done a little bit of a study into that. So one of the reasons is they went outside the search area, so the search area never reached them. A second reason was a team was assigned but didn't go to where they were assigned. That's pretty rare in this day and age with GPS.

Another reason is while they're conducting the search, a split occurs in the area, so they didn't cover that particular chunk of territory. And the final reason is it can just be darn hard to spot people out in the woods sometimes. So all it can take is a second or two of looking to your left when you needed to be looking to your right to spot the subject. Especially with children, they can crawl into small tight places that are obscured from view, so they can be very difficult to find. 

What does your app and your research … tell searchers to pay attention to when it's kids? What's different when you're looking for kids?

Probably that the hallmark with kids is even if you're shouting their name, they may not shout back because maybe they've been told not to speak to strangers, or they're just afraid of you, or they have an active imagination and they've turned you into Bigfoot or something.

They may think they're in trouble. So they may actively hide from searchers. And certainly we point out the fact that kids will crawl into thick underbrush. It may look impenetrable to you as the adult searcher standing up at five feet. But if you're down at between one and three feet, you may see a way to scramble underneath the brush and crawl into that. And that thick brush may provide … a certain amount of shelter to the child. So all those places needs to be searched. And that's a very difficult, time-consuming search.

How does your app help the searchers? 

Well, it gives them some of those distances of how far out they should be looking … It gives you percentage chance that they'll go uphill or downhill. They'll give you … kind of a checklist of where you should be searching first, where you should be putting your priorities.

It will give the actual field searchers a tactical briefing, it's like this is what you should be looking for. Thick brush and briars are important. Hiding places are important. The subject may not answer back to you, so don't rely upon that. Those kind of details would be in a field briefing that's in the app.

And just last thing before I let you go, is there anything you would want to say about this specific search?

My heart always goes out to the family and my heart also goes out to searchers who really pour everything they have into a search, especially a search for missing children.