From Ocean Navigator #107 July/August 2000 |
A video still of the M/V World Discoverer aground in the Solomon Islands. Was the accident caused by an SA-corrupted GPS position or was it due to an inaccurate chart?
On the morning of the next day, May 1, in Washington, D.C., 123° of longitude to the east, the White House announced President Clinton’s decision to remove selective availability (SA) from the civilian GPS signal. The President directed the Air Force to turn off SA at 8 p.m. eastern daylight time on Monday, May 1. This executive decision ends years of purposeful degradation of civilian GPS by the Department of Defense. With the flip of a switch it upgrades the accuracy of all GPS receiversunquestionably good news for ocean voyagers. With SA turned off, mariners can expect roughly 12-meter accuracy from a standard GPS receiver anywhere in the world.
Perhaps if M/V World Discoverer had steamed down Sandfly Passage in the Solomons Islands on Tuesday, with access to the full accuracy of GPS, instead of Sunday, it wouldn’t have hit the reef. Was the accident simply a case rotten luck? Before we get carried away, let’s look bit more closely at the World Discoverer accident. Would a more accurate GPS position have helped the ship to avoid grounding on the reef? According to an Associated Press story on the event, Michael Lomax, president of Society Expeditions, which runs World Discoverer out of Seattle, said the ship struck an uncharted coral reef. Perhaps the reef was, in fact, on the chart, but in order to save face, Society Expeditions insisted it was not. At this point we don’t know, but an uncharted reef is not at all outside the realm of possibility. The data on which some charts depend was gathered in the 18th or 19th century. The data on these charts doesn’t match a 12-meter GPS standard of accuracy. Even had World Discoverer been able to use GPS with SA off, it might still have struck the reef because the reef may not have been on the chart. Or perhaps the coral that forms the reef grew in the 200 years since the area was last surveyed and is now a hazard to ships.
The point here is that, while GPS accuracy of 12 meters is a wonderful thing in the abstract, in the real world of boats, water, rocks, and reefs, charts need to be accurate. While most of the charts for the waters of the developed world are reasonably correct, charts for many other areas on the globe sometimes are not.
So while the shutoff of SA is something to celebrate, voyaging sailors, who by their nature tend to sail to out-of-the-way places, should continue to exercise caution when in unfamiliar waters.
One GPS veteran who has lobbied for SA’s removal is a radionavigation consultant named Tom Stansell. Formerly a top engineer and technologist at GPS manufacturer Magnavox and later Leica, Stansell is an acknowledged expert in both the older Transit satnav system and its successor, GPS. When asked about the implications of the President’s SA decision, Stansell remarked that the dumping of SA will be a shot in the arm for the future of GPS. “I think it’s going to stimulate interest and further the development of GPS,” Stansell said. “I couldn’t wait to get out in front of my house with my handheld GPS and test it.”
Not only is GPS sans SA good right now, but it could get even better. According to a paper by Rob Conley and John W. Lavrakis, presented last fall at the Institute of Navigation’s GPS ’99 conference, civilian GPS has real potential to become more accurate than 10 meters. Conley and Lavrakis, both analysts at Overlook Systems Technologies, wrote that three major factors will continue to make GPS more accurate: improved accuracy in the atomic clocks carried aboard GPS satellites, better predictability of satellite orbits and clock states, and continuing the DOD’s excellent track record of keeping a full constellation of working satellites in orbit.
With all this superb accuracy and with the promise of better performance in the future, what happens to the Coast Guard’s differential GPS beacon network? Built in the 1990s by the Coast Guard specifically to remove the errors imposed on the civilian signal by DOD, the DGPS beacon network provides DGPS coverage on the East, Gulf and West coasts, the Great Lakes, many inland rivers, Hawaii, and much of the south coast of Alaska.
As of now, a few days after the White House announcement, the removal of SA has no effect on Coast Guard DGPS operation. “There is no impact on Coast Guard radionavigation programs,” said Cmdr. Curt Dubay, radionavigation program manager at USCG headquarters in Washington. “We’re going to continue to operate our marine DGPS system. And we’re going ahead with our nationwide DGPS expansion.”
The nationwide DGPS (NDGPS) expansion will provide differential corrections to most of the continental U.S. This increased coverage is part of the Department of Transportation’s Positive Train Control Project that seeks to eliminate train collisions and also allow railroads to better track shipments. In addition, the NDGPS expansion will provide coverage to all land vehicles. For mariners who use DGPS receivers, a bonus of the NDGPS system will be greater overlapping coverage in coastal areas. This translates into added redundancy. In some areas, should a DGPS transmitter fail, there will be another beacon within easy range.
While GPS with SA off is more accurate, DGPS promises still better accuracy than straight GPS. Selective availability was clearly the major source of error in the civilian GPS signal. The next largest source of error is ionospheric refraction of the GPS signal. The DGPS process removes ionospheric effects, further refining position accuracy. The actual accuracy level depends on a variety of factors, such as the type of receiver being used, the range of the user from the beacon station, etc. However, according to the Coast Guard, DGPS with SA off will be capable of one- to three-meter accuracy.
Of course, for recreational voyagers, this added accuracy is impressive, but somehow seems like overkill. (Winston Churchill reportedly said that after a certain point more atomic bombs did nothing more than “make the rubble bounce.”) After all, few recreational mariners really need a positioning source with an accuracy better than 12 meters. Okay, maybe there will be an occasion when you find yourself in pea soup fog and need to bring your boat into its berth using GPS alone. For the most part, however, even great positioning accuracy won’t (or at least it shouldn’t) prod you into that kind of risky behavior.
Turning off SA increases the accuracy of GPS, but it doesn’t do much about an issue that should give navigators pause: reliance on a single system for navigation. “The added accuracy to GPS is very good,” said Linn Roth, president of Locus, Inc., a manufacturer of loran and GPS equipment and the president of the International Loran Association. “But the whole issue of having a back-up system is still there.”
In U.S. waters the loran system is still operating, giving us two complementary systems, each with its strengths and weaknesses. Loran transmissions can also be used to send DGPS corrections. One intriguing differential GPS system, called Eurofix, was recently developed in Europe. Eurofix has a number of attractive advantages. The foremost benefit is that by using Eurofix, mariners have loran as a very capable back-up system for navigation should GPS become unavailable. Loran has proven itself a highly reliable radionavigation system with decades of service under its belt.
Without SA, differential users don’t require a high-speed stream of differential corrections. This is because, with SA off, the main source of error results when signals travel through the ionosphere. A low-data-rate signal like Eurofix is feasible because ionospheric errors don’t change much over the course of a few minutes.
A second advantage of loran is its range. The present loran network would be capable of sending corrections covering U.S. waters out to 600 miles. And a Eurofix system would have excellent redundancy. In most places in the coverage area, users would have multiple loran stations available should one go off the air.
Finally, loran and GPS can each be used to update each other. Way back in the dim mists of time1986 or soTrimble Navigation made a receiver called the 10X. This unit used loran and GPS in a complementary way. A modern version of this approach, combined with Eurofix corrections, could be an impressive navigation system.