When GPS was first being put into service, the US military was concerned about the possibility of enemy forces using the globally available GPS signals to guide their own weapon systems. Originally, the government thought the "coarse acquisition" (C/A) signal would give only about accuracy, but with improved receiver designs, the actual accuracy was . Starting in March 1990, to avoid providing such unexpected accuracy, the C/A signal transmitted on the L1 frequency () was deliberately degraded by offsetting its clock signal by a random amount, equivalent to about of distance. This technique, known as
Selective Availability, or SA for short, seriously degraded the usefulness of the GPS signal for non-military users. More accurate guidance was possible for users of dual-frequency GPS receivers which also received the L2 frequency (), but the L2 transmission, intended for military use, was encrypted and was available only to authorized users with the decryption keys. This presented a problem for civilian users who relied upon ground-based
radio navigation systems such as
LORAN,
VOR and
NDB systems costing millions of dollars each year to maintain. The advent of a
global navigation satellite system (GNSS) could provide greatly improved accuracy and performance at a fraction of the cost. The accuracy inherent in the SA however, was too poor to make this realistic. The military received multiple requests from the
Federal Aviation Administration (FAA),
United States Coast Guard (USCG) and
United States Department of Transportation (DOT) to set SA aside to enable civilian use of GNSS, but remained steadfast in its objection on grounds of security. Throughout the early to mid 1980s, a number of agencies worked to develop a solution to the SA "problem". Since the SA signal was changed slowly, the effect of its offset on positioning was relatively fixed – that is, if the offset was "100 meters to the east", that offset would be true over a relatively wide area. This suggested that broadcasting this offset to local GPS receivers could eliminate the effects of SA, resulting in measurements closer to GPS's theoretical performance, around . Additionally, another major source of errors in a GPS fix is due to transmission delays in the
ionosphere, which could also be measured and corrected for in the broadcast. This offered an improvement to about accuracy, more than enough for most civilian needs. The US Coast Guard was one of the more aggressive proponents of the DGPS, experimenting with the system on an ever-wider basis throughout the late 1980s and early 1990s. These signals are broadcast on marine
longwave frequencies, which could be received on existing
radiotelephones and fed into suitably equipped GPS receivers. Almost all major GPS vendors offered units with DGPS inputs, not only for the USCG signals, but also aviation units on either
VHF or commercial
AM radio bands. "Production quality" DGPS signals began to be sent out on a limited basis in 1996, and the network was rapidly expanded to cover most US ports of call, as well as the
Saint Lawrence Seaway in partnership with the
Canadian Coast Guard. Plans were put into place to expand the system across the US, but this would not be easy. The quality of the DGPS corrections generally fell with distance, and large transmitters capable of covering large areas tend to cluster near cities. This meant that lower-population areas, notably in the midwest and Alaska, would have little coverage by ground-based GPS. As of November 2013 the USCG's national DGPS consisted of 85 broadcast sites which provide dual coverage to almost the entire US coastline and inland navigable waterways including Alaska, Hawaii, and Puerto Rico. In addition the system provided single or dual coverage to a majority of the inland portion of United States. Instead, the FAA (and others) started studying broadcasting the signals across the entire hemisphere from communications satellites in geostationary orbit. This led to the
Wide Area Augmentation System (WAAS) and similar systems, although these are generally not referred to as DGPS, or alternatively, "wide-area DGPS". WAAS offers accuracy similar to the USCG's ground-based DGPS networks, and there has been some argument that the latter will be turned off as WAAS becomes fully operational. By the mid-1990s it was clear that the SA system was no longer useful in its intended role. DGPS would render it ineffective over the US, where it was considered most needed. Additionally, during the
Gulf War of 1990–1991 SA had been temporarily turned off because Allied troops were using commercial GPS receivers. This showed that leaving SA turned off could be useful to the United States. In 2000, an
executive order by
President Bill Clinton turned it off permanently. Nevertheless, by this point DGPS had evolved into a system for providing more accuracy than even a non-SA GPS signal could provide on its own. There are several other sources of error which share the same characteristics as SA in that they are the same over large areas and for "reasonable" amounts of time. These include the ionospheric effects mentioned earlier, as well as errors in the satellite position ephemeris data and
clock drift on the satellites. Depending on the amount of data being sent in the DGPS correction signal, correcting for these effects can reduce the error significantly, the best implementations offering accuracies of under . In addition to continued deployments of the USCG and FAA sponsored systems, a number of vendors have created commercial DGPS services, selling their signal (or receivers for it) to users who require better accuracy than the nominal 15 meters GPS offers. Almost all commercial GPS units, even hand-held units, now offer DGPS data inputs, and many also support WAAS directly. To some degree, a form of DGPS is now a natural part of most GPS operations. ==Stations==