New Mapping Tool…

SUBMITTED September, 2018
by JOE


My intrigue with the Fenn treasure began earlier this year. As soon as I read about it, my wife and I were hooked. As much as I wanted to know where the treasure was, it seemed like such an impossible feat trying to put all of the clues from the poem (as well as from Forrest himself) together into pin-pointing a location.

I did some research and found out there wasn’t really anything available to accomplish this. I’m employed as a Software Engineer and decided to take this project on myself. I began by implementing the items I knew were definite clues…. things like the possible states, elevation, and vegetation. Forrest also stated he only walked “less than a few miles”  and that searchers have been within 500 and 200 feet of the treasure.

As you can see from the map, I created a Google Maps search tool as well as a “Features” search. The Google Maps option is for searching specific areas, and the Feature tool is used for searching things more specific to the poem or hunches you may have (ie.: Brown, warm water, ect.).

With the layer selector on the left of the map, you’re able to choose one or more layers based on elevation, vegetation, and my “Trails” and “Roads” methods. Each selected tool has it’s own color in the map when checked.

As far as the layers go, here’s some basic explanations:
Elevation: (Almost) Everything between 5,000-10,200 ft. in elevation. I excluded Tribal lands from the beginning since I didn’t believe Forrest would hide it there, let alone encourage searchers to trespass. I thought it’d be pretty disrespectful.

Vegetation: I added a 500 ft. buffer to everywhere there was Sage and Pine (and Pinyon in one of the options since people have opinions about it), and found out where there intersected inside the correct elevation.

Trails method: I took every trail that USGS had a record of, added a 500 ft. buffer on each side of it, and then removed a 300 ft. buffer on each side internally, and clipped out the areas that didn’t match the vegetation/elevation layers. So you are left with a 200 ft. sliver on each side (giving the 500/200ft search area).

Roads method: I added a 3 mile buffer to every road in the “Treasure” states and clipped out the areas that didn’t fit the vegetation/elevation layers.

I worked hard on this map but also had a lot of fun with it. Most of the work was finding the data from USGS, parsing out the relevant information, combining and clipping the data from other sources, and finally making it easy to use as a tool. In the beginning I was just hammering my personal development laptop trying processing all the data. I would put ice packs in zip-lock bags under it so it wouldn’t thermal throttle when I would run my scripts that would sometimes take days. I ended up picking up a used rackmount server with dual Xeon CPUs and upgraded it to have 72GB of RAM. It’s been running non-stop for a few months now with different experiments I want to try.

Earlier this month my wife and I packed up our kids and headed on our own hunt based on information we gathered on possible locations from the map. Although we didn’t find the treasure, we found some wonderful off the beaten path areas that we wouldn’t have known existed without this map. We continue our search using the map and look forward to finding more search areas for next time!

My goal with creating this is having a reliable tool that every searcher can use. I plan to implement a social crowd-sourcing solves feature, and a few other neat features in the next few weeks (check the FAQ on the homepage for more info).

Feel free to email me any questions at:
Joe at ipson dot me
or if you’re feeling generous, Buy Me a Coffee

Good luck on your search!





35 thoughts on “New Mapping Tool…

  1. Pretty excited to see this! I was also trying to find a map online that would allow searching between specific elevations and such. Your software expertise is incredible…

    • Yes! More evidence of some truly amazing Americans on this site. While it does not solve my Ethical Dilemma it adds a lot of great exploring and a little Coffee to my Sunny-side-up solve.

  2. I am not understanding the vegetation Layer. If you look at a winter/fall view in Google earth, any place you see green in the rocky mountains, that is a pine forest of some kind. But more importantly, where you don’t see green, it may have been a pine forest in the recent past, or maybe will be in the near future, (in the 5000′ to 10,200′) zone. The reason ff picked 10,200′ (imo) is because in general/and on average, that is the tree line, this varies several hundred feet with micro climates of a given area, but in general that is a good estimate of the average tree line.

    I am not sure of the source you used for the vegetation layer, but it appears to not actually be the Pine forests of the rockies. What you could do is instead take a fall/winter view satellite image, and do a magic wand of anything that is green, and that would be a much better representation. although you could probably get almost as good an image in summer, since the rockies have nothing but pine forests of some kind, mixed with Aspen/Ash forests which are easily visible.

    To make it really simple, Satellite images give a very quick idea of the forests without needing a layer at all. so you could use just the paths with some percent of transparency, and the user could easily see where the forests are anyway through the transparency.

  3. Nice work. I have been working on a very similar sort of map in QGIS, based on at least a dozen different USGS and other data sources. I also wrote custom code, like you did, to detect geographical features that I interpret from the clues in the poem and from various things that Forrest has said. Unfortunately I now have about a million search areas to consider 🙂

    Through this process it has astounded me just how slow so many GIS routines are (due to being written in Python, and/or due to using algorithms that scale as O(N^2), which gets ridiculous once you have loaded all the roads or all the streams into the map for four whole states…). I’m not surprised you ended up upgrading to a Xeon box! I’m running on a 16-thread Ryzen processor, but that doesn’t help much when the GIS code is not even multithreaded!

    • I started with QGIS, and it was great to get my feet wet with GIS, but it is too slow to use for large calculations like this. I ended up writing my own buffer/dissolve scripts and using some other software to do the other operations. Which ended up being much faster.

      I was able to get around the lack of multithreading and just using python to fire off the same set of operations for every state, so I can actually use my cores.

  4. WRITIS,
    I am not trying to be
    contentious, because as you said, there
    are climatic conditions that can vary “Tree Line”. However, the major player is latitude.
    IN the Collegiate Ramge in Colorado, (basically
    N of Salida to almost Leadville, the tree line is
    a lot closer to 11,800 to 12,000 ft., where weather and geology permit.

  5. way cool Joe!

    Is it possible to make the layers subtractive? i.e. remove all areas not 5,000-10,000 and not within a trail area and not with pine/pinon, for example.

    macarena cool!

  6. Sounds like a nice tool – especially for those seeking a new spot. or someone just starting out. I am stuck on my area, and know almost every inch of it haha. If I did know every inch I would have found Indulgence, or I would have given up on the area – 🙂

    Thanks for your work – JDA

    • Why would that burst my bubble? Arc can do a lot, but the pricing is out of price range for most people who don’t use GIS professionally. This tool is for everyone to use for free. Arc would have been useless for creating the feature search anyways.

  7. One cool addition would be if you could also filter by land type (BLM, National Forest, etc.) as you might be able to snip off some areas of private property. I don’t know enough about the data gathering for this info to determine if it’s feasible.


  8. Cool. I wrote a similar tool with more search features, but I’m hesitant to throw my code up on github until after I go retrieve the chest. Your feature search needs some tweaking, it can’t seem to find Kismet Creek 😉

  9. I do applaud creative approaches to getting a leg up on the competition, while also mindful that Forrest doesn’t think “technology” will help you find the treasure chest. Of course, that doesn’t mean technology won’t be helpful in solving the clues! After all, Google Earth is technology; maps are technology; even books are technology. And everyone here is using a computer, tablet or smart phone, so technology is clearly integral to success.

  10. Did ArcGIS that was created years ago find the treasure? Nope, not at all. So why now use a different code with the same premise to searching. I believe this falls under what F said in SB 62.

    • CharlieM: I think the tool can be helpful for disabusing a searcher of a bad solution. Not everyone (especially tyros) are aware of all of Forrest’s ATFs.

  11. In addition to private vs. public and whether public is NFS, BLM, NPS, one final layer would be all the places concealing a 10″x10″ bronze box with gold and jewels. You can email the edition with that layer directly to me please. Thanks!

  12. I wonder how everyone here will spend their time if Forrest announces the chest has been found? Are people working other treasure search problems?

Leave a Reply

Your email address will not be published. Required fields are marked *