Most distant point

The most distant point from roads in the conterminous United States ("lower 48 states") is at the point indicated in the image below and in the file.



It shows the largest circle that can be drawn in the conterminous states that includes no roads or ocean. The circle radius is 21.7 miles, so the farthest that one can get from roads at a landlocked point is 21.7 miles. The point is close to the southeast corner of Yellowstone National Park and is inside the park, in the valley of the Yellowstone River upstream from Yellowstone Lake. The circle is bounded by U.S. Highway 20 on the north; South Fork (of the Shoshone River) Road on the east (connects to Highway 20); and Buffalo Valley Road on the south (connects to U.S. Highway 26). The red lines within Yellowstone National Park are trails. Coordinates are 44° 08′ 51.4″ N, 110° 04′ 25.3″ W.

Most distant point.kmz (1.91 KB)

Very interesting, Spacemaster.

You've got a pretty sizeable chunk of roadless real estate there, but please tell us how you can be sure that there is not a larger one lurking somewhere else.

I found this list which has the 50 most remote places and which locates the center of an 18.76 radius circle within about 5 miles SE of the one in spacemaster's post. Maybe some definitions* and more careful measurements have bumped up the size of  the West Thorofare Plateau site. 

*Peakbagger circles have to be entirely within Wilderness Areas and the other circle includes part of Yellowstone National Park, officially not a Wilderness Area, no matter how wild the part the circle includes. 

Rt. 16, the West Fork Highway, impinges just a little bit on the larger circle, but reducing the circle by that length of radius would still not endanger its No. 1 roadless status.

I had little doubt, spacemaster, that your assertion was correct. But any time a superlative is used, I think some explanation is needed about how that conclusion is reached. (OK, I won't fuss if you say your mother baked the best peanut butter cookies, but for most other things, I want to know how you got there.)

The other thought that occurs to me, is that roadlessness does not necessarily equal remoteness. (You, of course, never said it did, but I went looking in several areas before I realized that it would have to be in a place where roads were not permitted.) I thought about places like the Hell's Canyon area, or some places in Nevada and Southwest Oregon that have some roads, but would probably take much longer to reach, even with a 4x4 vehicle.

Both Washi  and Hill ask good questions.

Background

In 2006 and 2007 a team of geographers at U.S. Geological Survey (USGS) calculated a grid of the distance to/from the nearest road (DTR), on 30 m (100 ft) spacing across the conterminous 48 states. We published derivative results in the journal Science (I was lead author). The reference is: Roadless Space of the Conterminous United States
, Raymond D. Watts, Roger W. Compton, John H. McCammon, Carl L. Rich, Stewart M. Wright, Tom Owens, and Douglas S. Ouren
, Science 4 May 2007: 736-738. You can download this paper from the Science web site (you have to create a user name and password), or message me and I will email you a copy; I cannot post the paper on a public forum because of copyrights.

The analysis was done using Euclidean distance calculations on a grid in a GIS (Dijkstra's algorithm), in an Albers Equal Area projection. DTR grids were generated state by state, and for each target state a band of roads from adjoining states was included and then the results clipped to the target state's boundaries. Texas, Montana, and California were so large that they had to be calculated in segments and then the segments were joined.

Most distant points

We extracted the DTR values for each county in each state (a total of more than 3,000) and scanned these for the grid point with the largest DTR value, saving both the DTR value and the grid point coordinates. The most distant point from roads in the conterminous 48 states, according to our exhaustive calculations, is in Fort Jefferson in Dry Tortuga National Park off the Florida Keys, as shown below. Clearly, one can get right to this point in a power boat, and there are many other island and coastal points that are more distant from roads than the point in Yellowstone National Park, but they can all be approached by boat instead of road. Thus, the criterion of excluding both roads and ocean from a circle is crucial if one is interested in distance from machine supported approach. When one does this, the Yellowstone point is demonstrated as the most distant landlocked point.





Issues with grid analysis

There are problems with grid analyses based on Dijkstra's algorithm. It calculates distances accurately only along cardinal and diagonal directions. In essence, it approximates a circle with an octagon. Thus, when a "most distant point" is posted in a table, it is important to know whether the distance and "circle" center coordinate determinations were made with grid calculations or with a more accurate method. Even then, different results will be obtained depending on the projection used.

My offering of "most distant point"

The point that I posted at the start of this thread was determined by the most primitive of methods. I moved the Google Earth scene center around and drew circles of different radii, centered at the scene center, and adjusted the scene center and the circle radius until the circle just touched three roads (given three points that are not in a straight line, a circle can always be drawn through them). This method could be further refined by projecting the three road-touch points and then mathematically calculating the circle center. I knew where to begin, of course, from the earlier but less accurate grid analysis.

OK, spacemaster, you gave me what I ask for, in spades.

And now, I suppose, you're going to tell me that your mom won the Pillsbury Bakeoff.

Thanks for sharing your interesting and most impressive analysis. I'm eager to read more of your posts.

Yeah good point on the use of superlatives...

The tallest tree ? So someone checked them all ? It wasn't long ago that new species of trees were found, so obviously not every single tree stand has been checked for HEIGHT !

The fastest animal ? Does it count if its a mountain goat jumping into a river ?
A  gliding possum coming into land on a slippery slope ? it may  then travel as a land animal at an enormous rate...

One trivia question was "what is the only bird used for leather ? ".
Only ?? the heck ? Ok so we know that Australian emus are... Emu... No answer is ostrich ! .. Ah, hello you used an american trivia book, of course the emu is ALSO used for leather.. and like you can count the emu as the identical twin of an ostrich mate ! Nope, the answer is ostrich...

Isilder's apparent aversion to superlatives is okay with me. The Science paper referenced earlier in this thread went on to consider what the average distance to road was. This is a little bit like asking "what is the tallest forest," rather than "what is the tallest tree." What came from that analysis was quite interesting: the total space within some tract of interest (a county would be a good example) is better represented by the product (tract area) * (mean distance to the nearest road). The units of this measure are volume(!). There is value in learning to think and analyze in terms of distributions rather than extremes.

If you want to see a picture of how the volume measure of space makes sense, check the following web page:
Change in Roadless Space

A century ago, when great efforts were being made to reach remote points in polar regions and Africa, geographers identified "poles of inaccessibility." In the case of Antarctica, for instance, this was considered to be the point farthest from oceans. Wikipedia has a good starting point for exploring this topic, at Pole of Inaccessibility .

By the way, you are going up against Guinness!