Midway thru the expedition Trimble Navigation Ltd signed on as a sponsor and loaned us a complete Model 4800 Real Time Kenetic phase-differential GPS system plus a Field Representative to help set it up.

The First step was to find a location for the Differential Reference Station. The only practical spot was on top of the dive platform at the Wakulla Spring pool, which happens to be nearly over the entrance to the cave. This location was clear down to 10 degrees from the horizon in most directions, but the Reference Receiver occasionally lost a low angle satellite behind a tree for a few minutes. The Reference GPS Receiver was set up on a tripod over a mark on the concrete and turned ON to record raw data while we drove 10 miles away to set up the roving receiver at the nearest Florida State High Accuracy Reference Network (HARN) point. The benchmarks in or near the park were known to not be very accurate. Later, after some post processing, we  pinpointed the location of the Reference GPS to within a fraction of a cm. This point was the reference for all of our precision surveying.

The Trimble Reference GPS Receiver sent correction data continuously to a 25 watt UHF radio link with an antenna mounted on a 20 foot pole on the dive tower. This link worked everywhere in spite of the heavy forest cover which allowed us to do real time locations to within one centimeter.

The Trimble Roving Receiver was mounted on a carbon fiber vertical pole with a built-in bubble level and thumb-adjustable support legs. This setup was far easier to align over a point than the conventional tripod we used for the Reference receiver.  Once set up over a point, the Receiver had to pick 5 satellites in a reasonable pattern, then phase-lock onto both the civilian and military frequencies of each (thats 10 signals!), then calculate exactly which cycle of the carrier from each satellite it was locked onto. Multipath (signals that arrive with a slight time delay) caused by the trees often prevented a fix. Another problem was the apparently weak military signals, which had about 10 dB poorer signal to noise ratio than the civilian signals on these receivers. A separate data collector monitored the Receiver's progress and stored the the data from each fix.

    Outside the Park in the tall scrub growth, I would set up the Receiver then start cutting the small trees in an ever increasing circle until the Receiver got its fix. There was nothing more than 12-15 feet tall or more than two inches in diameter.  This whole operation only took one person (me) who became quite good at sharpening a machette. If it didn't work, I would come back on another day when the satellite pattern was different.  Eventually, every Radiolocation outside the park was directly fixed to within one cm.

    Several Ground Zero points inside the park (in open areas) were also successfully fixed, but the rest were under the old growth forest  canopy which caused too much multipath or simply blocked the signals altogether.  In the end, a total of 22 Radiolocated points were directly located with the Trimble gear.   When I had extra time I would go out with a helper, who would stand at a Ground Zero point while I located an open area where I could still see him, then used the Trimble Receiver to create a Reference point.  After I  created a second reference point,  we would clear the sight line. In this way, most points in the forest could be surveyed with a single shot using the Leica TC-1100 Total Station even though they were spread out over 5 square km.

    After each Trimble field session the data collector was connected to a laptop computer running Trimble Survey Office which downloaded the data and plotted the points.  Each Ground Zero point was compared to the approximate DGPS location as a reality check. Elevation data was collected for both the Trimble and Leica.