A six-channel Global Positioning System was used to locate eleven poin
ts on wetland boundary lines. The latitude and longitude of these poin
ts were compared to locations obtained by a second order ground survey
. Operating criteria and tree foliage coverage varied over a 19-month
period of data collection (February 1992 to October 1993). Canopy cove
rage levels sampled within a hurricane impacted area, did not increase
horizontal position error as much as percent of readings that had onl
y 3 satellite (2D) coverage. If at least 60 readings were made at each
point and more than 50% of the total readings had 4 satellite (3D) co
verage, the average horizontal distance error was 3m. Beginning in Jun
e 1993, with appropriate adjustments in data collection and data manip
ulation, 5-m accuracy was obtained 95% of the time regardless of time
of year, level of canopy coverage, or height of antenna. These accurac
ies seem to be sufficient for boundary surveys of delineated wetlands
for routine regulatory purposes. Criteria for data collection should b
e specified so that the accuracy of the data can be evaluated.