HOW FAR
IS IT?
By R. E. Lochbaum
Photo Credit:R. E. Lockbaum
How
far is it? That's a question I pondered
many times while hiking the various trails in the Smoky Mountains. It is also a question I was asked many times
by hikers I met while using a calibrated wheel to measure these trails. That
question is not easy to answer since it requires a careful definition of where
“it” is; i.e., from where to where. Another question I have been asked is why
did I measure them and how was it done? That question is easier to answer. The
desire to measure the trails in the Smoky Mountains evolved naturally from my
hiking activities. Walking has long
been a favorite leisure time activity for me. Upon early retirement in 1991 the
existence of free time allowed me to pursue some serious hiking. Joining the
Smoky Mountains Hiking Club was a step in that direction. As with many members
of the club, hiking every trail in the park soon became an objective. I teamed
up with member John Sloan, who had a similar hiking appetite and we completed
all of the trails on August 8, 1993.
Completion
of this goal did not end our desire for continued hiking and we mutually agreed
that having an objective made planning hikes easier. To this end we set a new goal to do every trail again in less
than a year. We had noted during the
first time through that several of the published trail distances, on maps and
signs, didn't seem right; therefore, I decided to collect information on these
distances during this second time through.
I
set up a data base for accumulating distance data. Using the 1994 edition of
the "Great Smoky Mountains Trail Map" as the reference, I defined
trail segments as the distance between trail intersections or between trailhead
and an intersection. Each trail segment and intersection was assigned a unique
designation. I planned to accurately measure the time to do each segment. With
apologies for the lack of humility, I have a steady and quite predictable pace
which permits reasonable distance estimates by measuring elapsed time. John provided me with an alternate backup
measurement by using a pedometer. This instrument had shown good results in
places where the distance was already known. In order to compare these measured
distances with those on the trail signs, a data base was created to record the
contents of each sign. I noted elapsed times, pedometer readings, and the
information on every trail sign with a tape recorder. All of the trail segments
were measured in this manner in nine months; thus, we met our “less than a
year” original objective. A review of
the resulting data bases did in fact reveal there were a number of trail
distances which appeared to be at odds with trail signs and/or map data. My fetish for detail which may result from
my German heritage required a more precise evaluation. Coupled with the fact we
wanted a new objective for hike planning, we decided to do each trail
again. This time I planned to measure
each trail segment with a calibrated wheel.
The
data base for collecting wheel data already existed. A measuring wheel was
fashioned using a rejected bicycle front fork obtained from a local bicycle
shop and the front wheel from my bicycle.
A handle was made using PVC piping and fittings. A counter was mounted for counting wheel
revolutions. About this time I become
acquainted with Steve Kemp of the National History Association who is the
motivator behind the little brown trails book Hiking Trails of The Smokies,
published by that organization. He
acknowledged several of the trail profiles in the book were lacking and new
measurements would be appreciated. To accommodate this an altimeter was mounted
on the wheel assembly. This setup,
which would also go through some evolution, was the device for starting a more
detailed trail measurement. As an Adopt-a-Trail
Volunteer in the Park (VIP program), I learned that backcountry camp pictures
were lacking so I added photographing each to the agenda.
The
first step in collecting the wheel measurements was to calibrate the
wheel. I was aware of survey markings
on the Little River Road extending from Sugarland Visitors center to Townsend
"Wye". These included an
accurately placed mark every hundred feet. (Historical Notation: These were lost when the road was washed out
by the flood and subsequently repaved in 1994). In my opinion the longer the
calibration measurement the better the accuracy; therefore, I ran the wheel
over several miles of the marks and recorded the distance and number of turns.
This also provided a shake down for the wheel assembly. Dividing the distance by the number of turns
gave the initial distance per turn calibration. I used this value for the initial segment measurements.
After
twenty hikes using the initial setup, the need for adjustment became apparent.
I used the wheel from my bicycle with an inflation pressure of 60 psi. This was fine for a trail with good tread (Hannah
Mountain), but became difficult on rough and rocky trails (Clingman’s Dome
Bypass). The high pressure caused the
tire to have a resilience somewhere between a super ball and a golf ball. The wheel was hard to control because of its
tendency to bounce. To solve this
problem I lowered the pressure to the point where it was firm enough to hold
its shape but not so firm as to be bouncy. As a result of this change I needed
a new calibration procedure. At the
high pressure the tire is stretched against its structural cords. A change in
pressure resulted in no noticeable change in circumference since above the
pressure it was fully expanded. At low
pressure a change in pressure did affect circumference. To deal with this, a fine piece of twine was
tied around the tire cross section to use as an index. It was rolled over a
smooth flat surface for three turns and carefully measured to within a 1/16th
inch; thus, for a single turn it was calibrated to 1/48th inch. The scientific
basis for using three turns was that’s all that would fit in my garage! After the first hike using the low pressure
configuration, I decided it would be better to have the same circumference each
time out rather than have it vary. A parametric study was run to plot wheel
circumference, or turns per mile, versus tire pressure. This led to the selection of a value that
resulted in an even number of turns per mile.
Since the pressure could not be accurately determined, the wheel was
calibrated using the three turn measurement to 1/16th inch method before each
hike. The next wheel change occurred when Steve Kemp gave me a second wheel.
This allowed me to put my bicycle back together again! This second wheel was a nominal 27 inch
wheel compared to the original 26 inch one. A new turns per mile versus tire
pressure parametric plot was developed for it and the three turn calibration
before each hike was continued. The
final wheel variation resulted in an even number of turns per 500 feet instead
of the even number for a mile. Since an altimeter reading was being taken every
500 feet, this simplified data collection. The majority (78%) of the trail mileage was measured using this final
setup.
During
the early phase of this project the need for an indication of precision was
recognized. To this end I decided to
measure each trail segment twice and to be acceptable, the two measurements
would have to be within 1/2 % of each other, which is 26 feet per mile. In cases where this standard was not met,
the segment would be remeasured until it was.
Careful attention was given to defining the end points of individual
segments so that the same length was being measured each time. Although the spacing of the wheels spokes
permitted recording measurements in the realm of an inch, to be practical
distances were recorded to one half turn of the wheel or about 3 1/2 feet.
To
obtain consistent altimeter readings it was necessary to eliminate errors
caused by parallax. This phenomenon is
the result of getting slightly different readings by looking at the instrument
from slightly different angles. This
can be mitigated by assuring that it is consistently viewed from the same
angle. To accomplish this. a system
analogous to the open sights on a rifle was added to the rig. Altimeter
readings were nominally recorded every 500 feet along the trail. With unique topography such as a prominent
peak or dip, intermediate recordings were made. On fairly level or very gentle grades where changes in elevation were
slight, readings could be extended to longer intervals.
A
typical data collecting hike included the aforementioned distance measurements
as well as the altimeter readings for profile development. The data was compiled using a small tape
recorder. In addition to trail segment distances, the distances to significant
landmarks were recorded. These included
stream crossings, notable artifact features, prominent geographical features, etc. Each numbered backcountry camp and shelter
was photographed. Usually each site was
photographed from three different angles.
Each trail sign was checked and where necessary previously recorded
information was revised and updated.
The data thus amassed was accumulated into several different data
bases. After 179 hikes which resulted
in measurements totaling 3271.4 miles, sufficient data had been collected. Data evaluation and manipulation was an on
going activity.
In
analyzing the distance measurement, the 1/48th inch calibration tolerance was
checked. This resulted in a maximum
error of 15 inches per mile or 0.025%. Since the wheel circumference was
sensitive to tire pressure, two sources of pressure change were examined;
namely, elevation and temperature changes.
Since there were significant elevation changes on many trails and since
ambient air pressure varied with altitude these effectively changed the tire
pressure. The results of this evaluation indicated that there was a 5 inch per
mile change in distance measured for a 1000 foot change in elevation. Increase in tire temperature caused an
increase in tire pressure. If the
ambient air temperature increased during a hike and it was assumed the tire
experienced the same change, there would also be a change in circumference.
From an analysis of this effect it was determined that a 10 degree change in
temperature results in a 4 inch per mile change in distance measurement.
Another source of error was running over
obstacles in the trail. Three categories were examined; namely, a round object
such as a limb laying on the trail, a round object such as a water bar half
buried across the trail, and finally a step - such as a large flat rock. From the equations for each case a series of
parametric curves were plotted. Three
variations of the round objects were examined. First was the case where the
wheel was rolled over the obstacle. In the second case, the wheel was rolled up
to contact the obstacle and then picked up and set on the other side and in
contact with the obstacle. Finally, the
wheel was rolled up to contact the obstacle and then picked up and set on top
of the obstacle. These curves were not
linear but typical examples for an eight inch obstacle gave various
results. For a flat rock step, about a
three inch increment was added to the distance measured. Rolling over an eight inch round object,
resulted in a 5 1/2 inch increase while lifting the wheel over it resulted in a
one inch decrease in measurement. For the half buried round object, the effects
were a two inch increase and a 3 1/2 inch decrease respectively. In summary, the elevation and temperature
changes resulted in distance measured being less than the actual, while the
obstacles generally result in increases; thus, their effects would tend to work
against each other. Furthermore, their
effects were less than a 1/10 % so it was concluded there was not a significant
degradation of accuracy.
Elevation
accuracy was evaluated indirectly since I had no means to calibrate the
altimeter. The difference in elevation
measured between ends of a trail was compared to the difference for that
segment derived from the topographic map.
The ratio of these two values provided an insight to the quality of the
elevation measurement.. Also since each segment was wheeled at least twice the
comparison of the two profiles provided a further check. Repeatability of these profiles when plotted
on top of each other proved to be quite good. Also, the ratio of the elevation
differences measured versus the differences derived from the topographic maps
for each segment were generally within 5%. Each profile was adjusted such that
its end points were in agreement with the topographic map. This was done in a
manner that maintained the integrity of the profile. Since there were ups and downs along many trail segments, the
difference between their end points did not generally give an accurate
indication of the actual gross elevation gained when traversing the
segment. To obtain the gross elevation
gain for a trail segment a computer program was developed which sums all of the
upward increments encountered in proceeding along the segment. The program was
run for each trail segment to provide the gross elevation gain in both
directions.
To
wrap this project up a computer program was developed to use these results for
planning hikes. By inputting the trail segments to be hiked it accesses the
different data bases to print out an itinerary which includes the trail
segments in sequence with the length of each, the starting and ending elevation
of each and the gross elevation gain for each.
The length and gross elevation gain are summed. A list of landmarks to be encountered along
the planned route is also provided. 
A printout of each trail sign on the route
presents the information as shown on the sign with the correct information in
parenthesis as necessary. A plot of the profile of each of the planned segments
is usually printed on back of the itinerary.
(See attachments 1 through 6.)
As
a conclusion the distances reported by the trail map and the trail signs are in
general quite good; however, there are differences and the following are some
interesting examples. By dwelling on these differences the impression may be
created that there are major discrepancies. There are many discrepancies but
few are major and usually measure about a tenth of a mile. For instance, on the
map there are less than two dozen cases where the difference is more than a
tenth of a mile and greater than a 10% error. Also there have been revisions to
signs and map. About one third of the mileages listed on the 1999 edition of
the map have been revised since the 1994 reference map. Approximately one
quarter of the segments shown on the map do not have their distances shown. One
principle discrepancy on the map is that the legend states distances are
between intersections while several are clearly the cumulative length of
several segments. There are several cases where the map and signs give
different lengths for a segment. The sign at the intersection of Forney Creek
and Jonas Creek trails would be a candidate for the least informative or most
disinformative. In addition to every item on it being incorrect it also has one
of the largest absolute errors in reporting the distance up to Forney Ridge as
being 6.1 miles when it is 7.4 miles. The mileage for a segment being different
on different signs occurs. The Hatcher Mountain Trail is an example. The sign
at its junction with Cooper Road says its 2.8 miles to Little Bottoms Trail
while the sign at Little Bottoms end says it 2.6 miles up to Cooper Road. The
signs along Deep Creek, Forney Creek, and in the area of the tunnel to “no
where” seem to have the most discrepancies. To reiterate, overall the signs and
map are reasonable.
A
frequently asked question is what is the total length of the park trails? To
give a precise response it is necessary to be specific as to exactly what is
included. For example since selecting the 1994 revision of the trail map as the
reference several changes have been made. There have been deletions (Hughes
Ridge lower end, Blanket Mountain, etc.), additions (first segment of Little
River), and rerouting of some segments (AT along Mollies Ridge). There are also
some ambiguities as to just what route segments take; for example, the Wet
Bottoms Trail has a variety of courses. Mindful of this, the approximate answer
is about 810 miles. Several other trivial facts can be extracted by
manipulating these data bases; for example, there are 392 trail signs in that
collection which includes some enroute navigationally informative signs. There
are 348 trail segments in the basic reference and thirty additional if the Wet
Bottoms variations and Cherokee Orchard area horse trails are included. These
were measured separately. The longest segment is Lakeshore from Hazel Creek to
Forney Creek at 21.030 miles. The shortest segment is the first part of the Old
Sugarlands Trail from the trailhead to the horse trail intersection at 141
feet. Three of the steepest sections are Hyatt Ridge, Low Gap, and Mount
Sterling all at over 800 feet per mile. There are 310 photographs of 106
campsites in the album. Five mini hand held tape recorders were used - three
being totally used up. Two and one half million revolutions of the wheel were
registered by the clicks of the counter that led to over 26,000 data elements
being recorded. The 0.5% precision objective was met with the overall average
for the over 800 miles of trails being 0.17% or less than 9 feet per mile. As a
follow up action since the completion of this project, the photographs have
been scanned into digital format.
During
a meeting with park personnel a comprehensive review of the data collected was
presented. All data including an album of the photographs plus a second set of
serial numbered and indexed photographs was volunteered. It is my understanding that the distance and
profile data will be used in a future revision of the Natural History
Association’s Hiking Trails of The
Smokies.
Also,
it was established that my hiking activities with John Sloan would definitely continue
toward our goal of completing every trail ten times. As an adjunct to this the
park’s Resource Management and Science Organization offered me the opportunity to
use a sophisticated Global Positioning System (GPS) to map the trails. This is
proceeding and as of December 1999 is over half completed. Under good
conditions the equipment is capable of accuracies of a meter or less. While the
desirable conditions do not always prevail, I believe accuracies of less than
ten meters are being attained with three meters or less being most common. To
put this in perspective, on a topographic 7 1/2 minute map with a scale of
1:24000, a line with the thickness of a piece of copy paper is seven and a half
feet (2.3 meter) wide. The accuracy achieved meets or exceeds the USGS mapping
standards for mapping with scales of 1:24000 or 1:12000. The instrument is set to record a position
every five seconds. The data provides a three-dimensional position for each
point. In addition to horizontal positions for mapping, the opportunity exists
to use the elevation data to revisit the profiles generated with the wheel
project data. This will be, time permitting, a future data reduction exercise.
How
far is it? As I said - it depends where “it” is. While I have a good grasp of
the distance for each of the elements to get “there”, I don’t really expect to
ever reach the terminus of my hiding adventures.
