The problem: why does a GPS receiver operated on the zero meridian at Greenwich indicate a longitude differing by about 100 metres from zero?
The basic longitude from a GPS receiver is referred to WGS 84 (World Geodetic System 84), which is a geocentric frame, with longitudes, latitudes and heights referred to a spheroid that best fits mean sea level over the whole globe. It is certainly true that the longitude of Greenwich is not zero in this system. It is also the case that the longitudes, latitudes and heights from this system will not agree particularly well with those from Ordnance Survey maps anywhere in the UK, and similarly in other countries it will be found that GPS coordinates will not agree well with the best maps of those countries. This is because within any country or region the maps are referred to a spheroid that best fits the sea level or mean land level of that region. This is necessary in order to give a sensible height system, in which heights are zero at the coast, and in which water flows downhill! So users of GPS should beware that they need to apply a transformation from WGS84 to the particular regional datum in order to be able to use the GPS positions on a map. As an example, at the Satellite Laser Ranger at Herstmonceux the WGS84 coordinates and the Ordnance Survey coordinates differ by 179, 65 and 45 metres in longitude, latitude and height. Many GPS receivers will give the option of outputting positions in one of these regional datums.
The Airy Transit Circle at Greenwich was originally the zero of longitude for the World. It is not at zero longitude in the WGS84 system, and the WGS84 system agrees with other modern systems (e.g. the International Terrestrial Reference Frame) to better than a metre. So at what stage was this origin lost from the International frames?
First it must be recognised that the only way that the position of any location can influence a terrestrial reference frame is if some positional measuring instrument is operated at the site, and the accuracy with which this position can be located in the frame is limited by the accuracy of the observations.
The latest astronomical observations made at Greenwich which contributed to the international terrestrial coordinate systems of that time were the observations made on the Small Transit instrument. These observations continued up to 1958. The observations made were the times and zenith angle of transit of stars, and were used to control Greenwich Time. Time signals from Greenwich were compared with time signals from other countries that were making similar observations, and from this a system of longitudes of the contributing observatories was obtained. The measured zenith angles give a system of latitudes. This coordinate system was compiled and maintained by the Bureau International de l'Heure (BIH). The zero point of this system was adjusted to give a constant value to the mean of all of them, rather than a precise value of zero for Greenwich, and so already at that time Greenwich had an offset of about 8 metres from the zero longitude of this system.
In 1957 the photographic zenith telescope (PZT) commenced operation at Herstmonceux. Efforts were made to transfer the longitude system from Greenwich to Herstmonceux, so that Herstmonceux could continue to maintain the same time scale. The transfer was in effect done by making observations at both sites for a while, and comparing the time scales generated by each, and adopting a longitude for Herstmonceux that gave the best match of the two time scales. In this way, after the observations from Greenwich ceases, the observations from Herstmonceux continued to give an indirect tie of the International frame to Greenwich. The accuracy of the transfer of longitude was probably at about the 10-20 metre level.
In the 1960s and 1970s Satellite methods began to take over as being the best methods to determine a terrestrial reference frame. Initially the tracking system was the measurement of range-rate from doppler. The attempt was made to orient these Doppler reference frames towards the Greenwich Mean Astronomical Meridian, as given by the BIH terrestrial frame. The Doppler technique was only accurate to the order of 10 metres, and so this imposes a limit on the accuracy with which the zero point can be set. Various geodetic frames set up using Doppler differ among themselves in origin by 10-20 metres. Another source of departure from the Greenwich longitude was that the Doppler datums were set up to have the same origin as the BIH system on the equator. The Doppler systems were reasonably close to being geocentric, but the BIH system was most likely far from geocentric. This may give some sort of skewness, so matching the frames on the equator would not give a good match at the latitude of Greenwich.
So all of these contributions may explain why the longitude of Greenwich in the Doppler systems would be the order of 100 m away from zero. Subsequent precise positioning systems, satellite laser ranging, VLBI and GPS, have based their longitude systems on that established by Doppler.