Maintaining topographic data
A printed topographic map makes geography look as if it naturally belongs to a sheet. Roads end at a neat border, contours disappear into the margin, place names are positioned for the reader, and the whole landscape is arranged at one chosen scale. For most of the twentieth century that was also close to the way national topographic production was organised. Survey, aerial photography, photogrammetry, field revision and cartographic compilation all fed the preparation and revision of individual map sheets. By the end of the 1980s, however, New Zealand’s national mapping organisation was beginning to maintain some of that geography as structured digital information that could continue across sheet edges and be used for more than one finished map.
The shift had a long prehistory. The first NZMS260 1:50,000 metric sheet, T12 Thames, appeared in 1977 on the New Zealand Map Grid, replacing the older one-inch NZMS1 series over time. Chapter 3 followed the parallel move into computer-assisted mapping, analytical photogrammetry and digital coordinate capture. Those developments made digital production possible, but they did not immediately turn the national topographic series into a database. A map could be plotted by computer while the authoritative production record still consisted mainly of cartographic drawings and sheet-based source material.
The change examined here was more fundamental. A road could become a database feature with geometry and attributes rather than only a line drawn on a particular sheet. The same feature could then be revised once and used in several outputs. A river did not need to be stored twice because it crossed a sheet boundary. Contours, vegetation, buildings and place names could be maintained as information in their own right, while the map became a cartographic view made from that information. That separation between geographic data and map portrayal became the organising idea behind the modern national topographic system.
A small-scale database first
New Zealand converted its national topographic archive in stages. The early national programme covered the 1:250,000 series. A 1989 article by and was already titled “The 1/250000 digital topographic data base of New Zealand: an outline”, placing a national digital database firmly inside professional cartographic practice by the end of the decade. A later technical account by , , and records that Surveyor-General approved a business case around 1987 to convert the 1:250,000 mapping into a digital database. The programme was expected to take about three years but was reported as completed in roughly eighteen months.
The conversion method reflected the technology available. The first work used a bureau service and SCITEX scanning and vectorisation rather than a modern desktop GIS workflow. Published map linework could be scanned, converted into vectors, structured and checked so that it behaved as connected geographic information rather than as an electronic tracing. Professional accounts describe the resulting database as seamless and topologically structured. Roads connected to roads, polygons closed properly and the database could understand adjacency and containment rather than storing only a collection of unrelated graphic elements.
GeoVision VISION was selected as the geographic database environment. That choice connected the topographic programme with the same broad generation of specialist workstation GIS described elsewhere in the book. The brand mattered less than the decision to create a maintained national information source rather than a set of digital drawings. Once the database existed, it could support map production, data supply and new forms of updating that were difficult when the paper sheet remained the basic unit of maintenance.
The database was also beginning to move outside the production office. By 1992, documentation for the New Zealand Land Resource Inventory referred to licensed digital topographic data at 1:250,000 obtained from the Department of Survey and Land Information. The national topographic database was therefore already useful as input to other GIS work rather than only as an internal cartographic tool. That external demand would become one of the forces pushing the more detailed 1:50,000 information toward digital form.
Converting the detailed map
The 1:50,000 programme was less tidy. The later technical history describes early conversion as demand driven, with priority given to particular areas and feature classes where customers or production needs justified the effort. Source material included original 1:25,000 compilations as well as the cartographic drawings prepared for the 1:50,000 NZMS260 series. This meant that the detailed national database grew through a mixture of programme work and practical demand rather than through one cleanly bounded conversion contract.
Terrain data helped force a more systematic approach. Demand for digital elevation information led to a formal programme to convert the contour source sheets nationally. The later account says that programme was completed within about twelve months and became an important test of the new VTRAK vectorisation capability acquired with an Intergraph Mapsetter 400 in 1993. The programme then extended across the other principal 1:50,000 map features. Contemporary United Nations reporting from 1991 had already described 1:50,000 capture as progressing rapidly, with national completion then expected by 1996.
The final three NZMS260 printed sheets were published in 1996. Individual digital conversion tasks began and ended at different dates. The printed-series completion and the development of the digital database were overlapping production histories. Some digital data existed while parts of the paper series were still being finished, and some database work continued after individual printed sheets had long been available. Treating 1996 as one magic conversion date would make a gradual programme look far neater than the evidence allows.
By the 1991 AURISA conference in Wellington, the professional question had shifted accordingly. Howard, O’Malley and presented “New Zealand’s Topographic Database - Where To Next?”, discussing the future use of the national database. The challenge was moving from conversion to maintenance, data modelling, distribution and repeated map production. Creating vectors from old linework was only the beginning of a permanent information-management task.
Conversion also forced decisions about classification. A paper map could use symbol, colour, line weight and context to tell an experienced reader that two similar-looking objects were different. A database needed explicit feature codes and attributes. The later programme developed a topographic data dictionary, drawing in part on the United States Spatial Data Transfer Standard, so that features could be classified consistently enough for exchange and repeated production. This was the less visible side of digitisation. A national database required agreement on both where a feature was and what sort of thing it was.
That problem became more difficult at 1:50,000 because the database contained a much richer landscape than the small-scale national overview. Tracks, roads, watercourses, vegetation, structures and named features created more combinations and more exceptions. A feature might change class without changing location. A road could be upgraded, a track abandoned, a building removed or a reserve renamed. The maintenance system therefore needed persistent identification, controlled classification and procedures for deciding when an external report was strong enough to alter the national record.
Data and cartography
A topographic database stores features and attributes; a topographic map presents selected information at a particular scale for a particular reader. A road may have one database centreline but be drawn with different widths or symbols. A building might remain in the source data while being omitted from a smaller-scale product. A place name has a geographic relationship to a feature, yet its printed position may be moved to avoid colliding with a road, contour or another label.
The digital system increasingly separated maintained source data from particular map products. Feature geometry, classification and attributes could be maintained independently from cartographic portrayal. Rules could automate much of the symbolisation, but cartographers still had to resolve difficult cases. A river crossing a road, closely spaced labels or a cluster of buildings could produce technically valid data and a terrible map. The database made consistency easier; it did not make judgement unnecessary.
There was also a positional legacy. Much of the early topographic database was populated by converting cartographic drawings that had been prepared to make readable maps, not to preserve the highest obtainable coordinate accuracy for every feature. Cartographers sometimes displaced features deliberately so that symbols could be distinguished. A road beside a river might be separated on the map even if their real-world positions were closer. Once that linework entered a database, the visual adjustment could look like an exact coordinate unless its provenance was understood.
Later maintenance increasingly sought better source information. Aerial and satellite imagery, photogrammetric work and authoritative data from other agencies could be used to improve geometry as well as keep features current. By the Topo50 period LINZ described using imagery and information from organisations such as the and the national road authority to maintain the source database. Digital production therefore changed the update workflow, but it did not eliminate observation. Someone still had to know that a road had opened, a track had disappeared, a coastline had changed or a place name had been amended.
Revision remained an information-gathering problem as well as a database-editing problem. Topographic agencies had always depended on aerial photography, field reports, other government organisations and local knowledge to find changes. Digital maintenance altered the destination of that evidence. Instead of marking up a drawing whose next life was another printed edition, staff could update a source feature that would feed several outputs. The potential efficiency was substantial, but only if change information arrived reliably and was assessed consistently.
That made relationships with other agencies more important. A national road authority might know about a new highway before the mapping agency did. Conservation staff might hold better information about tracks, huts and protected areas. Local authorities could know about urban expansion or renamed roads. Imagery could reveal physical change without explaining what an object should be called or how it should be classified. Maintaining the database meant combining observation with institutional evidence rather than assuming one source could describe the whole country.
The update cycle also changed expectations. A printed map edition could tolerate a lag because users understood that the sheet represented a particular revision date. A database supplied directly to GIS customers encouraged a stronger expectation that accepted changes should appear quickly. By 2007–08 LINZ was measuring how rapidly notified critical changes, errors and omissions appeared through NZTopoOnline. The production system was becoming a service with timeliness obligations as well as a cartographic archive.
People behind the database
The topographic database was built by a mixed workforce. Cartographers brought knowledge of generalisation, symbolisation and map readability. Photogrammetrists extracted terrain and feature information from aerial photography. Survey and geodetic staff maintained the coordinate framework. Database technicians, programmers and systems staff kept the specialist production environment operating, while conversion operators scanned, vectorised, coded and checked inherited linework. Field and agency information still had to be interpreted before it was accepted as a national update.
Technical papers associate , and with the database’s development and documentation. The 1999 account adds and during the move to Laser-Scan technology. Howard, and described another production change in 2009. The named specialists worked alongside cartographers, photogrammetrists, conversion operators and database staff who kept the national series running.
Topo50 generation required five cartographers to adjust text placement by hand. Experienced staff still checked how map-production rules handled contours, source-sheet differences and labels.
The government mapping and charting workforce included specialists whose names rarely appear in system histories. LINZ’s 2013 geospatial education material identifies Richard Freeman as a cartographer and Jennifer Ryan as Manager Chart Production, giving named practitioners behind the continuing transition of official map and chart production into digital workflows.
Sources · 1
A purchaser and a provider
The 1996 restructuring changed the organisation around the database while the digital transition was still under way. was split, with Land Information New Zealand retaining Crown responsibilities for national land information and taking substantial commercial mapping and production capability into a state-owned enterprise. For topographic mapping this created a purchaser-provider relationship around a national product whose authority remained with the Crown. Production expertise and contract work could sit outside the department even though LINZ remained responsible for the national mapping programme.
The 1998 Balclutha 1:50,000 sheet was published by LINZ and credits cartography to . That ordinary production credit records the division between the Crown organisation responsible for the official series and the commercial provider carrying out the cartographic work.
Technology was changing again at the same time. The earlier GeoVision environment became increasingly difficult to sustain, and the 1999 technical history records a reassessment after limitations in the system and changes in its supplier. By late 1995 the programme had selected Laser-Scan’s LAMPS2 object-oriented mapping environment. Implementation then overlapped the 1996 organisational split. Database migration, production software and government restructuring were therefore happening together rather than in a comfortable sequence.
The arrangement changed over time. During development of Topo50, LINZ later brought topographic data maintenance and database management back in-house. The same national dataset passed through internal government production, outsourced or contracted cartographic work, and later a more integrated in-house maintenance model as costs, technology and operational priorities changed.
In 1998 a user buying H46 received a conventional 1:50,000 topographic map on NZMG. It could be folded, carried, annotated and read without knowing anything about databases or government restructuring. Yet its production credit points to a digital and organisational system behind the familiar object. The visible map remained stable enough for users while the machinery that produced and maintained it was changing rapidly.
That separation between user experience and production history is easy to miss. A tramper could buy two editions a decade apart and see broadly the same cartographic language while the source data, software, contractor arrangements and revision workflow had all changed. National map series deliberately preserve continuity because users build habits and operational procedures around them. Digital production therefore had to modernise the source without making the product unnecessarily unfamiliar.
The map reaches the browser
Public access depended on distribution arrangements and licensing as well as digital production. For much of the 1990s, organisations wanting vector topographic data still dealt with specialist products, formats, licences and supply arrangements. A database could be national and digital while remaining largely invisible to an ordinary map user. The gradual appearance of topographic information online therefore represents a separate stage from the original database conversion.
By 2007 and 2008 LINZ was operating NZTopoOnline. Annual-report measures treated accepted critical topographic changes, errors and omissions as information that should appear through the service within working-day targets. Te Ara later described the national digital topographic database as available online through NZTopoOnline by 2008. This was an intermediate model: maintained government data had begun to escape the specialist production environment, but the later combination of free bulk downloads, open licensing, APIs and web services had not yet fully arrived.
The technical milestones occurred at different stages: capture, database maintenance, production and publication. First the map information was digitised. Then it became a maintained national database. Then parts of that information were distributed digitally to customers and displayed online. Later it became directly downloadable and, eventually, service-based infrastructure. Those stages overlapped, but they solved different problems.
A new grid for a new series
By the 2000s the old NZMS260 series also sat on an ageing national coordinate framework. NZMS260 had been built on the New Zealand Map Grid and the NZGD49 datum. Satellite positioning and the introduction of NZGD2000 created a more globally compatible reference system. LINZ adopted NZTM2000 for national topographic mapping in the early 2000s, aligning the map series more naturally with modern GPS and other national spatial data.
The projection change affected national mapping, data exchange and the alignment of existing spatial information. Sheet boundaries had to be redesigned. Existing users had coordinate references, indexes, emergency procedures and operational systems based on the old series. Police, Fire, Defence, the , ambulance, search and rescue and civil-defence organisations all relied on national topographic mapping in ways that made an uncoordinated switch risky. The new series therefore required data conversion, cartographic redesign and communication with users at the same time.
LINZ decided in 2005 to create what became Topo50, with a dedicated project team operating by 2006. The new 1:50,000 layout used 451 A1 portrait sheets, including six covering the Chatham Islands, rather than the 296 sheets of the outgoing NZMS260 printed series. A complementary Topo250 series used 31 sheets. Together the two series accounted for the 482 new maps recorded in LINZ’s 2009/10 annual reporting.
The Topo50 release
Topo50 was released nationally on 23 September 2009. The launch brought several long-running changes together in one visible event. The 451 1:50,000 sheets were released as a complete series, the old NZMS260 series was withdrawn, and emergency-service systems were shifted to the new framework. The maps used NZGD2000 and NZTM2000, and LINZ presented their compatibility with modern GPS and international mapping practice as a major practical improvement.
The physical maps were only half the release. TIFF and GeoTIFF images of all 451 Topo50 sheets were made available as free downloads, alongside digital topographic information that could be incorporated directly into GIS and other applications. The same national programme could therefore serve a tramper buying a printed sheet, an emergency service loading a map into an operational system and a GIS analyst who wanted the underlying digital information. Topography was no longer bound to one delivery form.
The production system behind those outputs had also changed. The 2009 technical paper describes a maintained national topographic database as the single source for the new maps. Cartographic rules generated most of the portrayal automatically, while operators intervened for difficult cases and text placement. Sheet-specific information such as marginalia and grid-reference examples could be assembled from definition files. The same workflow could produce PostScript, PDF, TIFF and GeoTIFF outputs, with map-generation times measured in minutes rather than months of manual compilation.
Printed sheets continued to be produced using automated processes. A paper map works without batteries, network coverage, software licences or a functioning server. It can be carried into the bush, spread across a table and annotated with a pencil. Emergency organisations continued to value nationally standardised printed mapping even as their GIS systems adopted the digital series. Paper maps remained in use after digital production was introduced. The database became the maintained source, and paper became one of several outputs generated from it.
The redesign created a large one-off cartographic job even though the source was digital. The 2009 technical paper describes automatic rules handling most feature portrayal, but text placement still required concentrated human work. Five cartographers worked through conflicts using raster versions of the old maps as one reference, moving labels and resolving cases that the automated process could not settle satisfactorily. Once those exceptions were stored, later production could reuse the result. Automation shifted effort toward designing rules and resolving exceptions rather than eliminating the initial labour.
Sheet marginalia became data-driven as well. Grid-reference examples, magnetic information, edition data and sheet-specific details could be held in definition files and inserted during production. Relief shading could be added from a separate raster source as part of the printing process. The map was increasingly assembled from controlled components rather than built as one indivisible graphic object.
The change also reduced the delay between database maintenance and new cartographic output. A feature could be corrected in the maintained source and carried into regenerated products without reconstructing an entire sheet by hand. The 2009 project described map-generation times of roughly ten to fifteen minutes for production files once the data and rules were ready. That did not mean a newly observed road reached a finished map in fifteen minutes. Verification, editing and quality assurance still came first. The speed applied to generating the cartographic product from an accepted database state.
From map series to spatial infrastructure
By 2009 New Zealand’s national topographic mapping system had crossed a threshold that had taken more than two decades to reach. The early national topographic database work had concentrated on the smaller-scale 1:250,000 series. Detailed 1:50,000 conversion then grew through demand, contour programmes and wider feature capture. Production systems changed from GeoVision to Laser-Scan technology, organisational responsibility moved through , LINZ and arrangements, and the source record shifted increasingly from cartographic drawings to maintained database features.
Cartographers increasingly applied their craft to rules, exceptions and portrayal rather than redrawing every feature for every edition. Photogrammetrists and imagery specialists continued to supply observations of the landscape, while database and GIS staff kept the structured source current. Map production became faster because repeated graphic work could be generated from maintained data, not because the country had become easier to observe.
The Topo50 launch made the new arrangement obvious to users. One national database supported printed sheets, downloadable images and reusable digital information on a coordinate framework compatible with modern positioning. The map sheet still mattered, but it was no longer the natural container of the national topographic record. New Zealand’s topography had become maintained spatial data from which maps could be made.
A parallel transformation was happening with a different kind of geography. Census boundaries and meshblocks were not designed to describe ridges, rivers or tracks. They divided the country so people, households and statistics could be counted, compared and linked to place. Once those boundaries also became maintained digital data, statistical geography became another national framework on which GIS users increasingly depended.
Chapter source notes
1. National 1:250,000 database. The period New Zealand Cartographic Journal account documents a national 1:250,000 digital topographic database by 1989. The later 1999 technical retrospective by directly involved practitioners documents the circa-1987 business-case approval, approximately eighteen-month conversion programme, scanning/vectorisation and VISION environment. This documents the programme history; an absolute national first is not established.
2. Detailed 1:50,000 conversion. The 1999 technical account documents the initially demand-led conversion, later national programmes, and the Mapsetter/VTRAK contour-conversion environment from 1993. Country reporting from 1991 provides corroboration.
3. LINZ and Terralink. The documented 1996 DOSLI split and the 1998 Balclutha 1:50,000 sheet, which records Terralink cartography and LINZ publication, document the purchaser-provider production arrangement.
4. Online topography. LINZ annual reporting and institutional history document NZTopoOnline being operational by 2007–08. The exact first-public launch date remains unverified.
5. Datum, projection and Topo50. Period LINZ geodetic material documents adoption of NZTM2000 for national mapping in 2001. Official launch material and contemporary technical reporting document national release of Topo50 and Topo250 on 23 September 2009, including 451 Topo50 and 31 Topo250 sheets. The LINZ 2009/10 annual report documents the combined 482-map programme and contemporary sources for free TIFF/GeoTIFF availability. Keep datum history in Chapter 19 and production automation in Chapter 43.