GPS Test Course—Puerto Rico

Introduction

To evaluate GPS receivers under a tropical canopy a Test course was established on the Caribbean National Forest (CNF) in January of 2004 under the direction of Carlos D. Rodríguez, ecologist with the USFS, International Institute of Tropical Forestry. The local Surveyor’s, Hector M. Sanabria Valentín, Luis Berrios Montes, and Hector Cruz performed all the surveying work. The Network was established from a baseline from two existing towers in the CNF, Youkahu and El Yunque. Mt. Britton, a third tower, has a control point with a horizontal order B (NGS Station AB9835). A new control point was established at the Yokahu Tower using three horizontal order B control points (Mt. Britton-AB9835; Fajardo-AB9834, and Humacao 2-AB9846). El Yunque tower has a second order control point (TV0917) that was reobserved along with the Yokahu Tower. The first two monuments of the test course have direct line of sight to the Yakahu Tower.

The CNF GPS Test Network is located in one of the most popular trails of the forest. The Big Tree trail originates at PR 191, km 10.4 in the CNF El Yunque Recreational Area. The asphalt paved trail is 1.4 km long and is a self-guided interpretive trail (signs in both Spanish and English are available along the trail), descending from 1,804 feet (550 meters) to 1,640 feet (500 meters). Hiking time is approximately 40 minutes, one-way. The trail is rated as moderate in difficulty, and is steep in some places. This trail is within the Tabonuco forest, the largest and most abundant of the four forest types found in the CNF.

Figure 1. The GPS monuments are located on the Big tree trail,
between road PR 191 and La Mima Falls.
The open site is located at the top of Yakahu Tower.

Figure 2. The Caribbean National Forest

The Tabonuco forest type is characterize by tall trees and low light intensities at ground level. Canopy trees tend to be very straight trunked with their first branches high above the ground. This forest type contains the richest flora on the Forest with 175 tree species. This zone appears on the foothills and slopes below 2,000 feet (610 meters) in elevation, which covers some 5,430 hectares (13,417 acres) of the Caribbean National Forest (USFS, http://www.fs.fed.us/r8/caribbean/natural-resources/). The mean annual rainfall range about 3470 mm for the past 30 years in the Tabonuco Forest (Neuman, 1994).

Figure 3. Typical vegetation around GPS monuments.

There are 7 monuments located along or near the trail that have been surveyed in. This site has a topographic obstruction to the west, a ridge, blocking satellites below 30 degrees at most of the monuments except monument number 1 and 2.

Figure 4. GPS Monuments located along the Big Tree Trail
with control and position brought in by traditional survey methods.

Base Station

The base station used to post-process the GPS data was the NGS CORS Base Station located at Isabela, PR, (PUR3) located in the NW corner of Puerto Rico about 137 km 277 degrees from the test course. On one of the days this site was non operational, so the Christiansted, St Croix NGS CORS site was used to post-process the data. It is located 139 km at a direction of 116 degrees from the test site.

Real-Time Beacon Station

For the real time beacon station, the Isabela PR site was used. It transmits on a frequency of 295 Mhz and at a transmit rate of 100 bps.

WAAS Satellite

The WAAS satellites are geosynchronous satellites in orbit near the equator. This satellite requires that the receiver be in line of sight. At the latitude of Puerto Rico these satellites are higher on the horizon and the correct signal more available that at latitudes of the northern US. At this location WAAS satellite PRN 122 was used to broadcast correction signals to the GPS receivers.

Test Procedure

The GPS receivers were set up on the monuments and 5 point and 60 point averages were taken for the different receivers and configurations. 3 repetitions were conducted for each receiver/configuration and then the average error calculated for the course and the RMS and NSSDA (95% confidence limit) values calculated.

TeraSync 3.31 software was used to collect data from the Trimble Geo-XM and Geo-XT receivers. Trimble Asset Surveyor software was used to collect data from the Trimble Pro-XR receiver. ArcPad 6.0.3 and SoloField 3.1.1 was used to collect data from the other GPS receivers tested. This data collection software was used to filter the positions collected so that only 3-D positions were collected with a PDOP less than 10. However the elevation mask (elimination of signals received from near the horizon as they are more susceptible to multipath or signal reflection), and SNR values (If signal strength is weak—more chance for error) are not filtered by these software packages and are subject to greater error possibility.

Observations

Due to the heavy canopy and the topographic obstruction to the west, sometimes long occupation times occurred before the required number of points was collected. This was especially true when only the internal GPS antenna was used. The external antenna increased the efficiency of the data collected. The less expensive receivers would use any signal received that met PDOP requirements but were weak or low on the horizon and a 60 position fix was obtained sooner but usually was not as accurate as the others that had other filter settings.

Results

The results of the GPS receivers tested are shown in table in Appendix B. A graph shown in figure 5 illustrates typical accuracies that can be expected with the tested receivers. This shows results with external antennas and 60 points averaged for these receivers. In field situations such as this, with topographical obstructions at certain horizontal elevations in known directions, mission planning can improve data collection efficiency by indicating times when satellites are visible in the open unobstructed skies and not when they are masked out by physical obstructions.

Appendix A

GPS Coordinates of Monuments

Appendix B

Receiver Results on PR Site

PDF version—best for printing (446 kb)