Frequently asked questions on surveying with the National GPS Network

Q. What is the difference between WGS84 and ETRS89?
Q. Why should I never use OSGB36^® triangulation points as GPS control stations?
Q. What is the best method for achieving high absolute accuracy in a GPS survey?
Q. Is the National Grid Transformation OSTN02^™ really as accurate as using OSGB36 control stations?
Q. Can I install my own height bench marks on the Newlyn datum?
Q. How can I obtain the best height accuracyfrom GPS?
Q. Where are the precise ETRS89 coordinates of the active stations?
Q. Do I need to take antenna phase centre offsets into account?
Q. What are the B, C, H and T prefixes in the Passive Station Number?
Q. What is RINEX format?
Q. How are the National GPS Network Active stations monitored?
Q. What is the MARKER NUMBER in the RINEX header?
Q. Why were the Active Network station coordinates updated in February 2002?
Q. How do I relate coordinates from the old and new transformation models?
Q. Why were the passive station coordinates changed in December 2002?
Q. Why were the Active Network station coordinates updated in September 2004?
Q. Why might I be having trouble reading Active station RINEX files into my GPS processing software?

Please tell us what you think about this document, or suggest a question

Back to the National GPS Network home page

Q. What is the difference between WGS84 and ETRS89?

A. Most people who are familiar with GPS have heard of the WGS84 (World Geodetic System 1984) coordinate system. This is a global coordinate system designed for use anywhere in the world. WGS84 coordinates are usually expressed as latitude, longitude and ellipsoid height.

WGS84 was designed for navigation applications, where the required accuracy is one metre or lower. A high-accuracy version of WGS84 known as ITRS (International Terrestrial Reference System) has been created in a number of versions since 1989, and this is suitable for international high-accuracy applications (it is used mostly by geoscientists). However, there is a problem with trying to use a global coordinate system for land surveying in a particular country or region. The problem is that the continents are constantly in motion with respect to each other, at rates of up to 12 centimetres per year. There are in reality no fixed points on Earth. In common with the rest of Europe, Great Britain is in motion with respect to the WGS84 coordinate system at a rate of about 2.5 centimetres per year. Over a decade, the WGS84 coordinates of any survey station in Britain change by a quarter of a metre due to this effect, which is unacceptable for precise survey purposes.

For this reason, the European Terrestrial Reference System 1989 (ETRS89) is used as the standard precise GPS coordinate system throughout Europe. ETRS89 is based on ITRS (the precise version of WGS84), except that it is tied to the European continent, and hence it is steadily moving away from the WGS84 coordinate system. In 2000, the difference between the ITRS (precise WGS84) coordinates of a point and the ETRS89 coordinates is about 25cm, and increasing by about 2.5 cm per year. The relationship between ITRS and ETRS89 is precisely defined at any point in time by a simple transformation published by the International Earth Rotation Service.

The ETRS89 coordinate reference system is used as a standard for precise GPS surveying throughout Europe. Using ETRS89 you can ignore the effects of continental motion: to a high degree of accuracy, the ETRS89 coordinates of a survey station stay fixed, as long as there is no local movement of the survey station. ETRS89 has been officially adopted as a standard coordinate system for precise GPS surveying by most national mapping agencies in Europe.

Q.  Why should I never use OSGB36 triangulation points as GPS control stations?

A.  Before the launch of this web site, there was a lack of precisely coordinated and monitored GPS reference stations in Great Britain, and it could be expensive to buy their coordinates. This situation led, understandably, to many surveyors using OSGB36 triangulation stations as GPS control points, obtaining GPS (WGS84) coordinates of these stations by transforming the OSGB36 archive coordinates.

The launch of this web site changes this situation entirely. Ordnance Survey now give you their entire National GPS Network of active and passive GPS control stations, at no cost, directly from this web site. We have done this because we are committed to encouraging best practice in GPS surveying throughout Great Britain.

It is poor survey practice to use an OSGB36 triangulation point as a GPS control station, because:

• A triangulation station does not have accurate GPS coordinates. All GPS surveys should be based on control stations with accurate and recent ETRS89 coordinates. No such coordinates exist for OSGB36 triangulation stations.
• A triangulation station is not maintained or monitored by Ordnance Survey for geodetic purposes. OS does not give any guarantees of the accuracy of archive OSGB36 coordinates, neither does it monitor the stability of OSGB36 monuments (with the exception of the clearly marked trig pillars which have been re used as passive GPS stations).
• Most OSGB36 coordinates were determined before 1950, and the system has never been overhauled.  The OSGB36 framework contains significant scale and orientation errors in a complex pattern, which can amount to 20 m relative errors over long distances. Why degrade and distort your high-precision GPS methods by using an outdated control framework?

For these reasons, only low-accuracy GPS coordinates (WGS84 coordinates) can be obtained from triangulation monuments. Therefore, although your resulting GPS survey may or may not have good internal consistency (that is, high relative accuracy), it will not be consistent with other GPS surveys in adjoining areas - that is, it will have low absolute accuracy. This effectively throws away one of the big advantages of GPS surveying, which is the ability to achieve both high relative accuracy and high absolute accuracy at little or no additional cost. The next section outlines how to achieve this.

Note: perhaps confusingly, some OSGB36 triangulation pillars have been reused as passive GPS stations, simply because they are convenient publicly-accessible monuments. These stations can be identified by a metal plaque bearing the words 'This monument forms part of the Ordnance Survey National GPS Network'.   The comments above do not apply to these stations.

Q.  What is the best method for achieving high absolute accuracy in a GPS survey?

A.  All GPS surveys should be based on accurate recent coordinates, in the ETRS89 coordinate system, of regularly monitored, positionally stable geodetic monuments. The active and passive stations of the OS National GPS Network are ideal for this, and are freely available from this web site. Our active stations are continuously monitored on a daily basis and have ETRS89 coordinates with (typically) 5 mm horizontal accuracy. Our passive stations are monitored on a five year cyclic programme, and have ETRS89 coordinates with (typically) 5 cm horizontal accuracy (better for some stations).

Using the National GPS Network reference stations, ETRS89 coordinates of your own primary survey stations, and hence of all your surveyed points, can be established using your usual GPS surveying methods. ETRS89 is the European standard precise GPS coordinate system, so your GPS survey will be consistent with thousands of other independent surveys throughout Europe. ETRS89 is in turn precisely related to all other precise GPS coordinate systems across the world, so your survey is ultimately tied in to the most accurate geodetic models of the whole earth. This is what we mean by absolute accuracy, and it has important practical benefits in terms of consistency and compatibility between datasets.

We recommend that you archive your survey coordinates in the form of ETRS89 latitude, longitude and ellipsoid height. If you need to convert these coordinates to British National Grid eastings and northings, this can be done to high precision using the Ordnance Survey national Grid Transformation OSTN02 (available on this web site in the Coordinate converter service). If you need to use a job-specific local mapping grid, you can do this in your usual software by defining appropriate map projection parameters. However, we recommend that you retain the archive of ETRS89 geodetic coordinates, so you can go back to them if required. If you need orthometric heights (mean sea level heights) of your survey stations, these are obtained using the Ordnance Survey National Geoid Model OSGM02 (available on this web site in the Coordinate converter service).

Q.  Is the National Grid Transformation OSTN02 really as accurate as using OSGB36 control stations?

A.  Let's take the National Grid Transformation OSTN02 first. This converts ETRS89 GPS coordinates to British National Grid coordinates, and vice versa. It is a complex transformation in that the parameters it applies vary in a complex way depending on your location in Great Britain. OSTN02 now defines the National Grid such that: ETRS89 (from National GPS Network) + OSTN02 = TRUE National Grid.

In discussing the accuracy of OSTN02, it's very important to understand exactly what we mean. The reason the OSTN02 transformation is needed is because the archived triangulation OSGB36, on which the National Grid is based, contains a complex pattern of distortions.

While it is true that two adjacent OSGB36 triangulation stations are usually in agreement with each other to within a few centimetres, if we compare two triangulation stations 10 km apart by measuring a precise GPS vector, we will typically find an error of several decimetres in their relative archive coordinates. If the two stations are 100 km apart, the relative error between them might be four metres. If the stations are at opposite ends of the country, we will find a twenty metre relative error between the two, based on their archive OSGB36 coordinates.

The OSTN02 transformation precisely models these distortions in OSGB36. So the first point to bear in mind is that converting a precise GPS survey to National Grid coordinates will always degrade its relative accuracy.  A rule of thumb guide to the maximum magnitude of this distortion is 4 centimetres per kilometre. If this distortion is unacceptable for your application,do not use National Grid coordinates. Instead, work in ETRS89 coordinates which have a greatly superior reference framework quality (less than 5 millimetres relative distortion across the whole of Great Britain).

National Grid coordinates should be used when the requirement of the survey has compatibility with Ordnance Survey mapping. The positional accuracy of Ordnance Survey large-scale mapping depends on the scale of mapping: the most accurate mapping is at 1:1250 scale. Detail features on this mapping are positioned to 0.5 metre accuracy (1 standard error) relative to the OSGB36 framework. Therefore, the transformation we use to convert GPS coordinates to National Grid coordinates must accomplish this conversion to better than half a metre accuracy.

The National Grid Transformation OSTN02 now defines the National Grid and models the old OSGB36 triangulation station framework to an accuracy of 0.1 m (1 standard error). This means it is five times more accurate than the most accurate Ordnance Survey mapping, relative to the OSGB36 framework. It is therefore more than sufficient for all survey tasks where the requirement is to achieve compatibility with Ordnance Survey mapping.

Q.  Can I install my own height bench marks on the Newlyn datum?

A. Yes. The services available on this web site allow you to compute high-accuracy orthometric heights (mean sea level heights) relative to the Ordnance Survey height datum without visiting any Ordnance Survey bench marks, and without knowing in advance the orthometric heights of the control stations you use. Ordnance Datum Newlyn (ODN) is the usual definition of mean sea level on the British mainland and some islands.

Any survey station with precise ETRS89 GPS coordinates determined using the National GPS Network can be used as a height bench mark by obtaining the equivalent orthometric height from the Ordnance Survey National Geoid Model OSGM02 (this is available from this web site in the Coordinate converter' service). It is advisable to use several such stations on the same site and level between them, to check the consistency of the orthometric heights. There is no need to occupy traditional Ordnance Survey bench marks by GPS, unless you have a specific requirement to check the relationship between your GPS survey and existing bench marks in the area (for example, if you have based previous height surveys on those bench marks and you need the new survey to fit exactly with the old).

However, be aware that obtaining precise heights by GPS is more difficult than obtaining horizontal coordinates. We recommend the use of IGS (International GPS Service) precise satellite orbits (ephemerides) rather than the RINEX navigation files supplied with the active station data (which contain the GPS broadcast satellite orbits). These are available from the IGS website. Mixing antenna types can effect the height accuracy of GPS surveys. It is important to tell your software the antenna phase centre offsets to apply. Read our advice on this subject here. The main practical implication is that GPS observation times must be longer to obtain precise heights than would be used for ordinary surveys. It is possible to obtain a GPS height with an accuracy of 2 cm (1 standard error) using the National GPS Network active stations, by recording four hours of continuous GPS observations at a primary survey station. Shorter observation times will generally produce less accurate results. In many cases, this will be more cost-effective than running levelling loops through a series of Ordnance Survey bench marks to establish Newlyn heights of your stations. And with the GPS method, you obtain the height of the station as it is today. Few Ordnance Survey bench marks have been checked more recently than the 1960s, so it is possible that their archive heights may in incorrect.

How accurate is the Ordnance Survey Geoid Model OSGM02? OSGM02 has an accuracy of 2 cm rms (1 standard error) for mainland Britain and better than 4 cms rms for other areas. This is a measure of the absolute accuracy or possible error introduced by the model. However, because the parameters which comprise the model change gradually, the relative error introduced by the model within a 10 km extent would at most be just a few millimetres, i.e. the applied shift will be almost the same across the area. This is a similar or better error profile to that obtained by high-quality national levelling, with the advantage that you do not have to rely on bench marks that may have last been heighted fifty years ago.

Q. How can I obtain the best height accuracy from GPS?

A. We recommend that the active stations are always used for high-accuracy heighting work, rather than the passive stations. The GPS observation period should be as long as possible: in our tests, typical 1.5 cm ellipsoid height accuracy (one standard error) was obtained from a 4 hour GPS observation, and using advanced scientific GPS processing software. To obtain the best possible heighting accuracy of about 1.0 cm (one standard error), careful site selection, geodetic quality GPS equipment, scientific analysis software, and a 24hr observation period are recommended. It is also important that antenna phase centre offsets are modelled correctly - see our advice on this subject here.

Q. Where are the precise ETRS89 coordinates of the active stations?

A. In the active station RINEX data files. The fields in the RINEX header marked 'approx coords' contain the official precise coordinates of the station. There is a comment to this effect in the RINEX header. The active stations are precisely monitored on a daily basis, and we may change these coordinates from time to time if we detect movement in a station.

Q. Do I need to take antenna phase centre offsets into account ?

A. For the best accuracy (particularly in height) - yes. Modelling the correct antenna phase centre offsets becomes important when using mixed antenna types. The Ordnance Survey National GPS Network contains a mix of antenna types. Full information on the antenna phase centre offsets in the Ordnance Survey National GPS Network can be found here.

Q. What are the B, C, H and T prefixes in the Passive Station Number?

A. The letter prefix in the Passive Station numbering system is largely a historical one indicating the origins of the station.

Stations prefixed with the letter 'B' are stations originally included in the SciNet (EUREF GB92) network of stations. Before the Active GPS Network these were of a higher order, however there is no longer any difference in quality between any of the designated Passive Layer stations.

Stations prefixed with the letter 'C' are stations which originally met the criteria for a National Network GPS (Passive Layer) station. Namely: have good aspect of the sky for GPS occupation; be accessible 24 hours a day to a two-wheel drive vehicle in all weather conditions; and permission for access, if required, easy to obtain. Most of these monuments were constructed specifically for GPS use, though where suitable, existing Fundamental Bench Mark or Triangulation monuments were adopted.

Stations prefixed with the letter 'H' or 'T' are Fundamental Bench mark or Triangulation monuments included in the Passive Network for transformation computation purposes, but do not generally meet the access requirements for a 'C' designated station.

The number immediately following these prefixes differentiates between stations when more than one exists within the same km square. Eg C2SZ3090, a station created in the km square SZ3090 to replace C1SZ3090 which has been destroyed.

Q. What is RINEX format?

A. RINEX stands for Receiver Independent Exchange and is a globally accepted standard format for GPS data.

Most GPS processing software will read in RINEX data as well as the manufacturer's proprietary binary format. This allows data from different manufacturer's receivers to be processed together which can be very difficult if data in only the proprietary binary format is available. Most GPS manufacturers should also supply a conversion utility to convert their binary format files to RINEX files.

All GPS data from the Ordnance Survey Active GPS Network stations is supplied in RINEX format to ensure maximum compatibility with all the different GPS processing software that users may have.

RINEX files usually come in pairs - a file of GPS observations (the "o" file) and a file of corresponding satellite positions (the "n" file). Generally GPS processing software will require both "o" and "n" files to be present in order to read in the RINEX data. There is a naming convention for RINEX files as follows: -

Q. How are the National GPS Network Active stations monitored?

A. The positional quality of every Active Network station is monitored on a regular basis. Daily coordinates, based on a whole 24 hours of GPS data, are computed for each station. The coordinate computation is carried out to the highest order using the “Bernese GPS Software” a specialist geodetic GPS processing tool.

The daily coordinates are compared with the station’s accepted coordinates (which are based on between 2 weeks and 6 months of data) and the differences in East, North and Up directions are computed. These differences are plotted on a graph to give an indication of the “positional noise” of a station. An example graph is given below for station THUR.