GPS TRAVERSE EXPERIMENT

Conducted by Ken Lucas & Paul Vincent

Objectives.
1. We wanted to determine if one or more of the available models of sportsman and sub-
meter grade GPS units could be used to determine the acreage of a forested area to an
accuracy equivalent to one gained by doing a conventional compass & chaining perimeter
traverse of the same area.
The difficulty in doing a perimeter traverse, with even a sub-meter mapping grade
GPS unit, has always been overcoming the error introduced to the satellite data from
“multipath” signal obstructions caused by overhanging forest canopy. We wanted to
determine if GPS data, collected at distant off-set points around the perimeter of a
forested unit, could be processed by GIS software to allow us to calculate an accurate
acreage for the unit of land.
2. We wanted to determine and compare the accuracy of a sportsman grade GPS unit to a
mapping grade GPS for determining point locations at both open off set points and points
under light to moderate forest canopies.

Equipment used.
• A trimble pathfinder mapping grade GPS unit. This is a older model sub-meter grade
gps unit. This unit was employed with the external antenna and supporting TDC-1 data
collector. This unit originally sold for $14,000 when new. This GPS unit allows
differential correction of the raw data.
• A Trimble Geo XT mapping grade GPS unit. This unit retails for approximately
$5,000. This unit was employed for these field tests without the optional external
“backpack” antenna. This GPS unit allows differential correction of the raw data.
• A Garmin 60CS sportsman grade GPS unit. This unit may be purchased on EBAY
used for approximately $300. This GPS unit does not allow differential correction of the
raw data.
• A LaserTech Impulse laser hypsometer and rangefinder.
• A Silva Ranger hand compass.
• ESRI GIS Arcview software
• Traverse PC software

Test preparations.
We completed conventional compass and chaining perimeter traverses around each of
the units used in these tests. These conventional traverses served as comparison
“controls” for the GPS traverses done during the tests. The Laser Tech Impulse
rangefinder and a silva hand compass were the tools used for these traverses. The bearing
and distance data was then processed in the Traverse PC software to determine the
acreage of the unit. All of the compass & chain traverses closed to under 1% error.
Field Tests.
The unit of land selected to function as the control for the first GPS perimeter traverse
test is a roughly rectangular 24-acre area, adjacent to a flood control levee along the
lower Eel River in the Redwood forest region of northern California. This unit offered
really clear lines-of-sight to the 360-degree horizon and represented the least potential for
multi-path errors of any of the three areas in this series of tests.
We ran 2 tests with the Garmin unit, in which we collected 500 readings at each of the
four main turning points on the unit perimeter. We employed the instrument in these
tests with the “averaging” feature, rather than the “acreage” feature available from the
menu suite
.
Results.

Traverse Acreage Error vs Control

The unit of land selected for the second GPS traverse test is a square 5-acre area
located on a wide alluvial flat, next to the Eel River, in the center of the Redwood forest
region, along the coast in northern California. The hills surrounding the unit have ridge
top elevations ranging from several hundred to 1,500’. The unit was covered by a 30-year
old young growth Redwood stand, with a stand density of approximately 300 trees per
acre and an extremely dense overhead canopy. There are 2-lane access roads along both
the east and west edges of the unit. The north edge of the unit is adjacent to a large
meadow and the south edge of the unit is adjacent to a recently thinned timber stand that
had a moderately dense overhead canopy. The other two sides had tree lines within 100-
300’. Also, two of the four turning points around the perimeter of the unit had a number
of old railroad rolling stock parked nearby, that offered a lot of potential for local
magnetic attraction to our compass. We ran two traverse tests each of the Trimble
GeoXT and the Garmin around this unit.

Results.
The control acreage calculated from the laser traverse was 4.7 acres for this unit. Both
tests with the Trimble GeoXT resulted in acreage figures that were less than 1% off of
the control acreage. Both tests with the garmin resulted in acreage figures that were less
than 2% off of the control acreage.

The final GPS traverse test was done on an approximately 24-acre forested unit
located on a wide, low ridge top in the Freshwater area of Humboldt County, California.
The unit had a very tall forest canopy, with a standing timber volume exceeding 65 MBF
per acre.
The compass & chain traverse consisted of 42 specific shots, for a total perimeter
distance of 5,491’. We achieved closure error of 14’ or 0.25%. As with the other field
tests, the acreage calculated from this traverse was used as a control figure to compare to
the GPS traverses. In addition, we used segments of the perimeter traverse to tie to 3 of
the 10 main turning points of the unit boundary; these three points were completely
obstructed by heavy forest canopy. We were then able to calculate a single bearing &
distance to each of these 3 turning points, which was then used to set the actual points
from the nearest GPS off set point.
We then proceeded to collect GPS data at each of the 8 points around the perimeter of
the unit both the Garmin model 60CS sportsman grade unit and a Trimble Pathfinder
mapping grade unit. For the Pathfinder, we collected data for 10 min or 100 data readings
with the PDOP Mask set at 6. For the Garmin we collected 500 data readings at each
point.
Two of the GPS off set points were in relatively small openings of approximately 75’ in
diameter, along the access roads in the otherwise heavy forest canopy.

Results.

Traverse Acreage Error vs Control

In our next series of tests, we wanted to compare the performance of the Garmin unit
to the older Trimble mapping-grade Pathfinder unit. The raw satellite data collected by
the Trimble Pathfinder was later differentially corrected and the raw data collected by the
Garmin unit was not differentially corrected.
The first test was simply a comparison between the two instruments at 12 points out in
the open. The locations of the points differed between the 2 instruments by as little as 3’
and as much as 44’, with the average difference being 23.5’ and a standard deviation of
11%.
The next test was to compare the two units in collecting point data under forest
canopies. In the several attempts we made, the Trimble Pathfinder was unable to collect
any satellite data at points under the various density canopies. We considered dropping
the PDOP mask on the Pathfinder to make the instrument accept lower quality satellite
data, but previous experience has shown us that the time to collect 500 points of data
could have required up to an hour of time. The Garmin, however, collected 500 points in
only a little more amount of time than both of the units would have required at points in
the open.
At 2 points under a fairly open canopy the Garmin point was off 57’ & 38’,
respectively, from the points determined from off-set points taken in the open.
At 3 points, under fairly heavy canopies, the error ranged from 39’ to 59’.
Finally, we tried using the Garmin to collect data in a very small opening in the forest
canopy(maybe 20’ in diameter) and determined an error of 36’. The Trimble Pathfinder
could not collect data in this small opening.

Summary points.
• GPS traverses appear to be a very viable method of determining acreages of units.
An error of 1%, as achieved in many of our field tests, means that a GPS traverse of a 24-
acre area might only be off by a quarter-acre.
• For determining acreages of units, the Garmin sportsman-grade GPS unit seems to
provide a really impressive accuracy.
• For determining locations of points under forest canopies, either from distant off-set
points out in the open and in nearby small openings, the Garmin unit can provide
satisfactory accuracy, especially considering the cost difference of several thousand
dollars between the different grades of instruments and the delay required to differentially
correct the mapping grade raw data.
California land managers regularly have to commission traverses on specific units of
timberland, in order to verify the acreage of these units. An accurate acreage is needed in
order to support various land management activities, which commonly include
completing timber cruises to determine the value of standing timber and verifying that
government-mandated harvest acreage limits are adhered to. In addition, a land manager
may want to verify a discrepancy in the county’s assessed acres for a specific unit of
land, in order to later justify commissioning a more expensive licensed survey of a
property. Indeed, a licensed survey traverse of a property could run $65 per acre. Even a
less accurate compass & chain traverse of a property could cost $32 per acre. So, a land
manager might have to spend $1,500 determining the acreage of a single 25-acre harvest
unit.
We realize the importance of being able to offer land managers a much cheaper
alternative to these conventional traverse methods. Our test verified that the GPS
traverse method could achieve an equivalent accuracy to the more expensive compass &
chaining method and cost only $13 per acre. This could mean that a GPS traverse of a
25-acre of land could only end up costing $250
Of course, we realize that there may often be situations where the nearest acceptable
satellite reception point (the off set point) might be further away from the unit edge than a
single traverse shot. In those cases where the off set point is a long way off, it might still
be necessary to employ the laser traverse method for a segment of the unit perimeter.
However, even if a third of a 25-acre unit required compass & chaining (with the other
two thirds permitting GPS traversing), because of a lack of nearby forest openings; the
cost would still be less than half the cost to laser traverse the entire unit perimeter.
Land managers, interested in saving time and money, should evaluate every unit-to-
be-traversed to determine a traverse plan that would allow the best blend of the GPS and
compass & chaining traverse methods.
Please feel free to email us at ken@tsiwood.com for more information. My website is
www.tsiwood.com