Mapping and coordinate systems ˜ an update
The following article first appeared in Newsletter 122:13-16 (August 2000) of the Geological Society of New Zealand, and is reproduced here with permission. Note that a link is given in the References to a related article in Survey Quarterly 1999.
Mapping and coordinate systems ˜ an update
Ian Reilly (ireilly_at_southern.co.nz)
Two years ago, I outlined a number of changes that were being proposed for our national geodetic datum and coordinate system (Reilly, 1998). Some of these changes have potential impacts on our national topographic mapping series at 1:50 000 and 1:250 000, and in consequence on specialised maps derived from these, such as the GNS QMAP series.
Land Information New Zealand (LINZ) has now promulgated the New Zealand Geodetic Datum 2000 (NZGD2000), a reference frame aligned with the International Terrestrial Reference Frame 1996 (ITRF96), and with its origin at the mass centre of the Earth as determined by satellite methods. This origin is displaced by about 200 metres from the origin of our existing New Zealand Geodetic Datum 1949 (NZGD49), on which all our current series of topographic maps are based. Moreover, NZGD2000 is provided with a geocentric Cartesian coordinate system, as well as a geodetic coordinate system (latitude and longitude) based on the reference ellipsoid defined by the Geodetic Reference System 1980 (GRS80). The latter ellipsoid differs in size and shape from the International (Hayford) Ellipsoid 1924, which defines the geodetic coordinates of NZGD49.
A new map projection?
That NZGD2000 and NZGD49 are associated with different ellipsoids means that a map projection for the one cannot be identical with a map projection for the other. This has prompted LINZ to open discussion on the need for a new map projection, and to hold a one-day Workshop in Wellington on 8 December 1999, at which I was privileged to be amongst the ca. 20 participants. The workshop was "facilitated", and the criteria so predetermined, as to lead to the conclusion, admirably summed up on p. 6 of the "Discussion Paper" produced by LINZ after the meeting:
"Under the Survey Act 1986 LINZ has a mandate to produce hard copy topographic maps for emergency management and Defence purposes. The existing NZMG map projection used for 1:50 000 ˜ 1:500 000 topographic mapping is explicitly tied to NZGD49 through the use of the same ellipsoid. To gain some of the potential benefits afforded through the adoption of NZGD2000, e.g. the increased accuracy and compliance with international standards, it is desirable to adopt a new map projection in terms of this datum."
with the conclusion that
"a standard projection in the form of Transverse Mercator could best satisfy New Zealand‚s current requirements and long term benefits"
This has the aim, inter alia, of covering the 24° range in longitude of the entire Exclusive Economic Zone (EEZ). Such a projection would replace the New Zealand Map Grid Projection (NZMG), on which are based all our current topographic map series on scales of 1:50 000 to 1:500 000, and its adoption would entail the forced obsolescence and replacement of these series.
The tenuous basis for the LINZ recommendation
I consider that most of the arguments adduced above, and elsewhere in the "Discussion Paper", in support of the single Transverse Mercator projection, are either false, unsubstantiated, or irrelevant, as follows:
That "increased accuracy" of positional data demands a new projection: false ˜ the projection has no relevance to the accuracy of positions displayed in it. Any well-defined mapping projection will transform coordinates virtually error-free, limited only by the rounding accuracy of the computational process. The accuracy of the positional data is an entirely separate question. LINZ has determined that positions referred to NZGD49 are up to 6 metres in error, and has produced a "distortion grid" ˜ a computational algorithm ˜ that is claimed to correct for this to an accuracy of about 0.2 metres. It should be noted that errors of 6 metres are insignificant at the 1:50 000 scale of the Topographic Map 260 series, though they are important for detailed mapping at the decimetre level, say. Positional data for 1:50 000 mapping has an accuracy of about 25 metres; it is totally misleading to imply that the use of a "new projection" will, by itself, enable such data to be safely used at a scale of 1:1000.
- That the projection "meets defence requirements": spurious ˜ "defence requirements" can be satisfied by a Universal Transverse Mercator (UTM) grid overprinted on any regular topographic map, including the present series on NZMG. A general Transverse Mercator projection has no particular advantage. New Zealand lies across three UTM zones (58, 59, 60), which are designed to each span 6° in longitude. The relevant UTM grids can be overprinted on the present Topographic Map 260 as required, either in place of, or in addition to, the NZMG grid.
- That the projection "meets international standards": unsubstantiated ˜ in spite of requests, LINZ has failed to identify these so-called "standards" of any general application. There are however, some specific requirements that were brought up at the meeting:
- The UN Cartographic Office requires a map of New Zealand on a conformal conic projection at a scale of 1:1 000 000; such a map has been in print for many years;
- The hydrographic service (now LINZ; formerly RNZN) continues to produce nautical charts on Mercator Projections at various scales, in accordance with international agreements, and harbour charts on an ad hoc basis on local Transverse Mercator projections;
- The civil aviation authorities produce aeronautical charts on stereographic projections, again in accordance with international requirements.
- That it is "GPS/GIS friendly": largely irrelevant, as accurate survey data are to be acquired in NZGD2000 formats, and archived as NZGD2000 coordinates. Moreover, while reputable suppliers of GPS equipment will provide read-outs in UTM and even NZMG, they may not cater for arbitrary TM coordinates
- That the single Transverse Mercator projection is needed to cover the EEZ: pure speculation, as the meeting produced no firm proposals for a seamless series of large-scale maps over the EEZ ˜ the only real justification for a single projection ˜ and it was conceded that future maps would likely be on various scales, projections, and origins.
I conclude that no convincing argument has been raised to justify the replacement of NZMG by a Transverse Mercator projection. For mapping the land area of New Zealand, and the contiguous seas, NZMG still has the advantage of a much smaller scale error than Transverse Mercator, and it is in widespread use (Reilly, 1999). Given the power of modern computers, the task of turning any coordinates ˜ NZGD49 or NZGD2000 ˜ into NZMG coordinates (or those of any other mapping projection) is trivial. That, however, is no excuse for the unnecessary proliferation of coordinate systems.
The threat to our mapping series
The present 1:50 000 Topographic Map 260 series was begun in the 1970s to supplant the former one-mile maps (1:63 360). The result was a series much improved in scale and resolution. Present LINZ proposals for a redesigned 1:50 000 topographic map series, on a new and different projection, offer no such improvement in data quality, just the forced obsolescence of the maps of the present series. Those who are in contact with the map-using public report the slowness with which users become acclimatised to new map scales and coordinate systems. Neither the general public nor the emergency services will benefit from a needless disruption to familiar systems.
Not every country that has introduced a new geocentric datum has so lightly proposed to cast off its existing topographic maps. The Swiss, with an outstanding reputation in topographic mapping, are keeping to the projection and coordinate system for topographic mapping that was adopted in 1903, and which is based on the Bessel Ellipsoid of 1841 (Schneider et al., 1999). The Swiss have also produced, however, an upgrade (LV95) to their original (LV03) national cadastral coordinate system, to remove distortions of up to 1.6 metres, and, alongside, have introduced a new geocentric system (CHTRF95) ˜ the analogue of NZGD2000 ˜ for fundamental record purposes. The latter has an origin displaced by ca. 800 metres from the centroid of the Bessel Ellipsoid.
New Zealand has even less reason ˜ a population density one-tenth of that of Switzerland ˜ to scrap its present high quality topographic maps, instead of continuing to update the existing sheets as necessary. Perhaps we need to define an NZGD96 ˜ as a modernised NZGD49 with the distortions removed, and an analogue to the Swiss LV95 ˜ to help to clear away the muddled thinking with which the LINZ proposals are riddled. I suspect that an "NZGD96" is already the de facto standard datum for GPS manufacturers who provide read-out in NZMG; they are hardly likely to put the NZGD49 distortions back into the coordinates.
Of course, those who produce specialised maps, such as the GNS QMAP series, can carry on using the same projection and grid as before. But the most serious threat is to the continuing existence of the topographic series on which most geological mapping is based.
A disturbing impression that I carried away from the workshop was of the reluctance of LINZ spokesmen, under direct questioning, to affirm any long-term commitment to maintaining topographic map series at scales of 1:50 000 to 1:500 000. In the light of this, the quotation given above from the LINZ "Discussion Paper":
"…LINZ has a mandate to produce hard copy topographic maps for emergency management and Defence purposes???
could imply that LINZ may be interpreting its mandate so narrowly as to exclude the interests of the scientific community and the public at large. This is ultimately a question of Government policy, and it would seem appropriate, if not urgent, for the Geological Society, along with others, to impress their support for continuing topographic mapping upon the relevant Ministers.
I thank John Beavan and Robin Phillips for their comments on the draft of this article.
Reilly, Ian, 1999: Which way forward? Change and choice in mapping New Zealand. (with an Appendix "The origins of the New Zealand Map Grid Projection"). Survey Quarterly 20:30-35.
Reilly, Ian, 1998: Mapping & geodetic coordinate systems: a background to change. Newsletter of the Geological Society of New Zealand 116: 13-17 (also New Zealand Geophysical Society Newsletter 50: 13-17; reprinted with additions in Survey Quarterly 16:14-17, Dec 1998)
Schneider, D, Gubler, E., Marti, U., Gurtner, W., 1999: Aufbau der neuen Landesvermessung der Schweiz ŒLV95??? Teil 3: Terrestrische Bezugssysteme und Bezugsrahmen. Bericht 8, Bundesamt für Landestopographie, Wabern.