Mapping the REWHC Trails
In order to create trailmaps reflecting the REWHC trail system, our mapping techniques are evolving. Rough sketches, overlayed on site plans, are giving way to more accurate mapping by surveyor's compass and measuring wheel. Ideally, differential GPS could be used for more accurate mapping, but the cost of implementing differential GPS and its accuracy benefits in our well-contained system do not justify use at this time. We will, however, explore ordinary GPS usage and whether the errors of using inexpensive handheld GPS devices provide suitable tracks for map making.

Initial Trail Mapping
We were fortunate to start with a good site drawing from Raytheon's Facilities Department. We were able to correlate trail landmarks with that of the drawing, and free-hand the rest. A measuring wheel was used to get accurate distances for the trail guide, but the free-hand sketch was not accurate to a scale.

Exploring New Mapping Techniques
The addition of the new Meadow Meander trail, brought an opportunity to experiment with a new technique; the use of a surveyor's compass and measuring wheel. Starting from a known location, a sequence of distances and angles from magnetic North are recorded at appropriate intervals and a vector map is created. The Meadow Meander, being a small, looping trail, is an ideal candidate for prototyping this mapping process. The nature of a loop trail, intersecting itself, allows accuracy of measurements to be known. The small size of the loop means that multiple techniques may be prototyped and accuracies evaluated for the eventual use on the remainder of the trail system.

Initial Wheel and Compass Results
The initial results were encouraging, with less than 11 ft of error in closing the loop. The following are the measured vectors taken from the light pole between the rightmost wildflower gardens while traversing the loop in the counter clockwise (CCW) direction:

Table 1: Data from Initial CCW Measurements

ID	A (deg)	R (ft)	(in)	. . .	ID	A (deg)	R (ft)	(in)
V1	63	50	7	. . .	V12	148	53	3
V2	206	50	7	. . .	V13	120	34	0
V3	122	115	9	. . .	V14	100	54	11
V4	55	246	10	. . .	V15	77	122	2
V5	45	151	11	. . .	V16	43	32	3
V6	134	48	9	. . .	V17	15	116	0
V7*	184	57	6	. . .	V18	344	56	7
V8	210	77	1	. . .	V19	305	160	7
V9	195	83	3	. . .	V20	343	80	7
V10	181	88	1	. . .	V21*	288	93	3
V11	161	125	4	. . .	. . .	. . .	. . .	. . .

*V₇ and V₂₁ are the loop's vertex points.

Entering the above data into a CAD package as vectors, the following image is drawn:


Figure 1: QuickCAD Rendering of the Table 1 Vectors



Figure 2: QuickCAD Measurement of the Closure Error



Figure 3: Spanning the Closure Error in the Last Vector

Discussion of Error
Sources of error include distance, angular measurement, and transcription. To apply this measurement technique with confidence over greater distances, these sources of errors must be known and minimized:

• Distance Measurement Error - Two flavors of distance measurement error are anticipated, one due to the wheel rolling over ground contours and the other due to roundoff errors at each vector. The use of a measuring wheel over uneven ground will likely be our largest concern, especially where the trail is not smooth, but has a lot of contours. If the contours are consistent across the trail, a factor may be calculated by measuring the distance between two points both with a tape and a measuring wheel and creating a correction factor. The Rolatape MM-30 is certified accurate within one inch per hundred feet on a smooth hard surface. On other surfaces, wheel slip may increase this error. The use of measurement tapes or chains is another way to minimize the distance error, but take more time to use. Round-off errors can not be avoided, but the fewer measurements used, the better for round-off purposes. The problem with fewer measurements, is the squaring off of what may be a curvy trail. While distances are more true, the trail will appear more angular than in reality.
• Angular Measurement Error - Two flavors of angular measurement error are anticipated, one due to the accuracy of the sighting process itself and one due to the angular roundoff accumulations. The Brunton SurveyMaster, allows sightings to within a half degree in experienced hands. It uses a technique where sighting is performed with both eyes open. A condition, known as heterophoria, must be checked on those providing bearings using the compass, since disalignment of the eye axis introduces additional errors. This is a simple check which can be performed by a person in the field. It may be best to have the same person perform the sighting for consistency. Fewer measurements would reduce accumulated round-off error, but at the price of curve-fitting fidelity.
• Transcription Errors - The most accurate measurement, if transposed during recording, provides another potential source of error. Since the person measuring and recording may be different, better accuracy is obtained if the recorder loudly and audibly repeats the called measurement for the measurer to hear.
  
  

Ultimately, we're talking about minimizing positional errors. Known positions on site maps can be used as datums to "anchor" estimated measurements. For example, additional bearings to a known or common location provide another measurement factor to reveal transposed data and improve positional accuracy. Multiple sets of data may create a "mean" path for trail purposes. Loops may be recorded both clockwise and counterclockwise and their paths averaged as well. Wherever possible, known, unmovable locations should be used in the data to anchor sets of mesurements in position.

Next Steps
The next step is to revisit the Meadow Meander to hone our surveying skills:

• The Meadow Meander will be traversed in the CW direction,
• A "mean contour" factor will be computed and used to adjust measured distances on the clockwise data set,
• The CW loop closure error will be computed,
• The "mean contour" factor will be applied to the original CCW data set and the loop closure error computed,
• The CW and CCW paths will be averaged and the mean path assessed for standard deviation and closure,
• The secondary bearings will be used, first adjusting for angular measurement at all points and computing loop closure error (assuming distance measurements accurate), then adjusting distance at all points and computing loop closure (assuming angular measurements accurate).
• The findings will be presented and analyzed for "best techniques" in moving forward with site mapping.

Update
We experimented with a Garmin Etrex Vista GPS and walked the trails on a cool February morning, logging position every five seconds and downloaded it to the Garmin MapSource program with Eastern US Topo Maps. Figure 4 depicts the output directly from the computer download from the GPS raw track file:

Figure 4: GPS Track Plot from Garmin Etrex Vista and MapSource

In order to add buildings and other such features, a topographic map, Figure 5, was referenced and correlated to the GPS Tracks.


Figure 5: The Topo Map for The Area Located Roads

Using photoshop, the routes and features were traced onto multiple layers, correlated, and annotated with feature names. An updated trailmap resulted as shown in Figure 6, and downloadable from here.


Figure 6: Tracing of Trails, Roads and Other Features on Map