 |
| http://geographicprofiler.com |
Geographic profiling is a relatively new (1996) tool in the serial crime investigation toolbox. There are MANY different variables that can be inputted into the various mathematical formulas and the more data points you can enter then, hypothetically, the more accurate the profile will be. In short, the formula helps to narrow down the possible location of a "base" for a serial offender. The formula utilizes data such as the crime scene location, dump site (if any), proximity to and types of transit available, locations where weapons might have been obtained, and possible locations where the offender might have picked their victim. It's very in depth but it's still just a tool. Like psychological profiling, geographic profiling cannot solve crimes by itself. But it has proven quite useful in maximizing resources for search areas. Can this tool be used to help identify a suspect for the Axe Murders? The map at the right utilizes Kim Rossmo's formula for geographic profiling. The map shows the geographic profile of Dennis Radar (BTK). The area in red is the predicted base of Radar and lays on top of the College Hill neighborhood of Wichita. The problem with geographic profiling becomes evident here. Radar lived in the north, in Park City, and was mobile. He would park his car away from the crime scene, then walk to and from the victims' houses. In four cases, he drove the victim's car back to his own vehicle. Radar also stated that he "trolled" for victims just by walking around places. I believe only one of his victims was within walking distance of his house. Radar's house didn't necessarily need to be considered his base. Four of his seven crimes occurred between 1974 and 1979, when he was attending Wichita State. Wichita State falls in a fairly cool spot on the search zone and is roughly two and a half miles from the College Hill neighborhood and roughly 9 miles away from his actual house.
So with that example in mind, can geographic profiling be applied to the Midwest Axe Murders? Unfortunately I don't think so. At least, I don't think I can get a truly accurate location using these tools. This map shows the seven (blue markers) crimes
I consider to be connected to the Midwest Axe Murders. The gray markers are possible crimes but I'm not certain of it. They are the Caseways of San Antonio who were killed in March of 1911 and Belle Jordan of Mount Pleasant, IA, who was attacked in her house Halloween morning in 1911. The distances traveled are enormous and the two most widely used geographic profiling methods (Rossmo's formula & Gaussian Rossmooth) are built around a series of assumptions, one of which, doesn't fit our crimes. "The criminal only resides in one location - By this, we mean that
though the criminal may change residence,
he or she will not move to
a completely different area and commit crimes there." In other words, neither system is designed for this level of mobility in an offender. Both models also assume a smaller, rectangular area that contains all crime scenes. This clearly doesn't fit our crime series as the area shaded in blue represents 62,000 square miles! However, if you look at the markers on the map you'll notice the nearly linear arrangement of the crime scenes. All of the crimes follow single lines of rail but if you draw the lines in chronological sequence you get the below map, which is much more convoluted. A person could ride the train from Portland, OR to
Tacoma, WA. This would take roughly 24 hours and pass through Rainier, WA. From Tacoma a person could get to Denver and switch trains to get to Colorado Springs. This leg would take a few days. A person could get from Colorado Springs to Monmouth via Kansas City or go back through Denver. If they took the second option, it would be about 855 miles with a lot of stops in between, including Villisca and
Mt. Pleasant. All of these locations fell on the Burlington Route of the Chicago, Burlington and Quincy Railway. Also along the Burlington were Paola and Ellsworth, although they were not passenger stops. To my knowledge, the CB and Q railroad was the only railroad that stopped or passed through each town in which one of these murders occurred. If you want to check for yourself, here's a
link.
What does any of this tell me? Well, maybe the killer wasn't just hopping random trains. The railroad system was so complex, I can't see some random person knowing where each and every train might be going. I think a person might be able to pick out the limited and express routes but not every bit of rolling stock that ran on the rails. The CB and Q was huge though. It owned most of the rail between Chicago and Denver so it could be a not-very-big coincidence. It also tells me the killer might have been familiar with these towns before the murders. I have to stress, if they rode from Denver to Monmouth on the CB and Q, they would have stopped in both Villisca and Mt. Pleasant on the way. I also want to be clear that I am not suggesting the killer was buying tickets. I am using the Railway Guide as a way to trace the routes of passenger trains not freight trains or work trains. The killer could definitely have been just jumping on and off trains.
2 comments:
Excellent post, as always. I'm pleased to see you're still going with the blog. The killer's mobility and use of the railroads would certainly seem to confound any conventional approach to geographic profiling. To your observations I would add only that I believe he was most likely 'based' (during the Midwestern phase of the murders - the last four in particular) in or near one of the major rail hubs in the area, from which he could travel directly to each of the crime scenes. The fact that the murders observe a somewhat radial rather than linear pattern suggest that he was striking out from a fairly fixed location rather than riding the rails cross-country as a drifter or transient. I don't see any reason why he would change trains or make multi-leg journeys (a necessity if he didn't live or work near a major junction) in order to travel to towns that appear to have been chosen at random - it seems to me that he was simply travelling outward in various directions from a 'hub' location and selecting a likely-looking town along the route. It's interesting that you point out an apparent (perhaps even exclusive) preference for the CB&Q Railroad. Is it possible, do you think, that he was an employee of that company?
I've just re-read your excellent post on geographic profiling, and I believe there are a few further points to consider. I have some background in geographic profiling for law enforcement here in the UK, and I hope these observations are of some interest:
You correctly point out that highly specific geographic profiling is rendered nigh on impossible for this type mobile and wide-ranging offender. This is indeed true to the extent that a pinpoint residential location (think Flower and Dean Street for Jack the Ripper) or even a likely neighbourhood (as with your reference to Dennis Rader) would be out of the question. However, it would not be at all beyond the realms of possibility to make informed suppositions about a city in which the unsub was resident, even given the vast distances involved. In general, the more mobile the suspect, the less precise the resulting location.
The roughly linear west-east pattern you correctly observe is a common illusion when viewing map data outside of its temporal context. If we are to accept that the same person was responsible for both the Pacific Northwest and Midwestern crimes (which I believe is probable but not certain to the extent that it should be used to rule possibilities in or out), profilers would divide them into two distinct phases - for ease of reference, let's call them phase 1 and phase 2 - for each of which a separate geographic profile would need to be compiled.
With only two crimes in phase 1, such a profile would likely not yield any results. Colorado Springs may have represented a transitional location on the long route between his phase 1 and phase 2 residences (more can be inferred from this, and I can go into detail if you wish). The four widely recognised phase 2 murders - Monmouth, Ellsworth, Paola and Villisca - observe a radial, crisscrossing, back-and-forth pattern in a (relatively) confined area. There is no discernible directional trend to suggest ongoing movement. He would appear, for this period, to have settled in one place.
Post a Comment