Digitising map sheets
By the late 1970s, the was using a computer-assisted system to convert cadastral mapping into digital form. A paper cadastral sheet was fixed to a large digitising table, control points were established, and an operator traced selected lines and points with an electronic cursor. The system recorded coordinates rather than an image of the page. Boundaries, roads and other features could then be stored, edited and plotted from the computer instead of being redrawn entirely by hand.
The system was MAPPAK. developed it during MSc work in Computer Science at the , using requirements supplied by Lands and Survey. Colley’s 1979 thesis, A Computer System for a Cadastral Mapping Application, described its use for rural and urban cadastral compilation. It covered automated conversion between map series, computer-produced cartography, graphical and descriptive data structures, editing and plotting, as well as proposed land-information functions. The distinction between operational and proposed features is retained in the thesis.
Computers were already being used for survey adjustment, engineering calculation and coordinate transformation. Those applications could accept geographic measurements, calculate results and print an answer without retaining a reusable map database. MAPPAK retained the linework itself in digital form. Once a feature had been captured, the same geometry could be corrected, transformed and plotted again without returning to the original sheet.
Metric conversion
MAPPAK developed during a period when Lands and Survey faced a large conversion programme. New Zealand was changing from imperial to metric measurement and adopting the New Zealand Map Grid for national metric mapping. published the NZMG design in 1973, and the new 1:50,000 NZMS260 topographic series began appearing later in the decade. The first sheet, T12 Thames, was published in 1977.
Metrication affected much more than labels and scale bars. Existing cadastral and topographic information had been compiled under older measurement systems, projections and sheet layouts. New metric series required coordinates, linework and annotation to be transformed or recompiled. Lands and Survey had thousands of sheets to deal with, so repeated mathematical conversion and automated plotting were obvious candidates for computer processing.
NZMG was designed specifically for New Zealand and used a polynomial formulation rather than one of the more common projection implementations later built into international software. It reduced scale distortion effectively across the country, but its specialised mathematics later caused compatibility problems in generic navigation and mapping software. During the 1970s the immediate task was simpler: convert existing material into the new national mapping framework and produce replacement maps. Colley’s thesis treated that conversion as both a cartographic and data-management problem.
The thesis separated several tasks that had often been part of one manual workflow. Existing linework had to be captured. Coordinates had to be transformed. The resulting geometry had to be checked and edited. Cartographic output still had to meet acceptable production standards, while descriptive information associated with the geometry had to be stored in a form that could be reused.
Inside MAPPAK
The MAPPAK installation used a DEC PDP-11/34 with 48K words of memory, two RK05 disk units and a TS03 nine-track magnetic tape drive. Input and interaction were provided through a DECwriter terminal, a Tektronix 4010 graphics display and a Summagraphics ID48 digitiser measuring 36 by 48 inches. Output went to a Broomall 430/101 flatbed plotter. Colley recorded the complete installation as costing less than $100,000. This small family of specialised machines had to cooperate before a line could travel from paper to screen and back to paper again.
The arrangement already contained the main components of a later digital mapping workstation, although spread across separate specialist devices. The digitiser captured coordinates from the paper map. The graphics terminal allowed the operator to inspect and edit the linework. Disks and tape stored the data, while the plotter turned the digital geometry back into hard-copy cartography. Software ran under the PDP-11 environment rather than through the graphical desktop interfaces that would become familiar later.
Operators digitised map features as coordinates. A scanner would reproduce the page as an image, including linework, text, symbols, marks and imperfections. A digitising operator selected the features to capture and traced them as coordinates. The process required decisions about which lines represented geographic objects, how features connected and what descriptive information needed to be retained.
Colley reported that a dense 1:2,000 cadastral base sheet containing substantial hydrographic detail could take about two hours to digitise. The operator had to register the source sheet correctly, select and trace features, inspect the result and correct errors. Two hours was a production rate for this supervised work. The computer reduced repeated redrawing while the operator remained responsible for checking.
The accuracy of the digital output remained tied to the source. A displaced or generalised line on a paper map did not become survey-accurate when converted to coordinates. Digitising could also introduce its own positional error. The computer could store more decimal places than the source justified, but the additional digits did not change the quality of the original mapping.
Stored geometry
Conventional cartographic production could create an accurate and well-drawn new sheet without preserving the features as independently reusable digital objects. MAPPAK separated data capture from final plotting. A line could be stored once, edited later and used in another output. Descriptive information could also be stored with the graphical data rather than remaining only as text drawn on the sheet.
This altered the role of the map sheet inside the production process. Paper sheets were still used as source material and remained the main form of distribution, but the computer file could hold geometry independently of one printed sheet. A road or boundary crossing a sheet edge did not have to stop at that edge in the stored data. The production system still used sheet structures, but the geometry itself could be maintained separately from the physical limits of the paper.
Colley also designed structures for graphical and descriptive information that anticipated broader land-information processing. The thesis discussed ways of interrogating and manipulating stored data. MAPPAK can be described securely as an operational computer-assisted cadastral mapping system with land-information-system characteristics. An absolute claim that it was New Zealand’s first GIS remains unsupported.
MAPPAK addressed cadastral conversion and digital storage before the later Digital Cadastral Database (DCDB). The two programmes tackled related problems but remained distinct.
Routes and terrain
The Ministry of Works was also applying computers to geographic problems during the 1970s. Its engineering work involved road and rail alignments, earthworks, gradients, terrain and hydro development, all of which generated large calculation workloads. Retrospective accounts describe computer-assisted route selection using information derived partly from aerial photography. The computer could evaluate more alternatives than an engineering team could reasonably calculate in detail by hand.
later described the Mangaweka road-rail deviation as one case in which computers generated or evaluated possible trail lines subject to engineering controls such as gradient and earthworks. In the Clutha Valley, ten road alternatives were reported as being examined through computer processing where manual methods might have allowed detailed consideration of only two. The computers supported engineering analysis.
These route programs used geographic measurements and engineering constraints for particular projects. They were designed to solve specific engineering problems. Later GIS software incorporated similar forms of terrain and route analysis inside broader environments, but the underlying calculations had already been used in specialist engineering systems.
Terrain representation was also being automated. In 1976 of the New Zealand Agricultural Engineering Institute published Contouring by Computer from Stadia Field Data. The program converted surveyed positions and elevations into contour information by interpolating between irregularly distributed field observations. Computer-generated contours are now routine, but the 1976 work records part of the shift from manual interpolation to software-based terrain processing in New Zealand.
A contour is derived rather than measured directly. Field observations provide positions and heights, and an interpolation method estimates the surface between them. Moving that process into software made the method repeatable and allowed larger sets of observations to be handled consistently. The resulting contour still depended on the quality and distribution of the source observations and on the interpolation method used.
Orthophoto production
Aerial photography had been part of New Zealand mapping for decades before computers entered the workflow. Photogrammetrists used overlapping photographs, ground control and stereoscopic instruments to measure terrain and compile map detail. By the 1970s, computers were being introduced into parts of this production chain, particularly where geometric correction and coordinate calculation could be automated.
Lands and Survey purchased a computer system for orthophoto production in 1977. Orthophotos correct much of the displacement produced by camera geometry and terrain so that the resulting image can be used within a map coordinate framework. The corrections require numerical information about the geometry of the photograph and the ground surface.
Orthophoto mapping was produced for the Clutha-Alexandra area during 1978 and 1979 for Ministry of Works hydro-development planning. The product remained recognisably photographic, but its geometry had been corrected through a computer-assisted production process. The imagery could then be compared more directly with mapped features and engineering information. This was still specialist mapping production rather than general-purpose GIS.
Analytical photogrammetry
followed another route into digital production. In 1984 the company acquired a Wild BC1 analytical stereoplotter, replacing part of the older optical and mechanical geometry with computer-supported calculation. The operator continued to view overlapping photographs stereoscopically and identify terrain and features, while the analytical instrument calculated the geometry of the stereo model. Measurements could be recorded as coordinates rather than only driving a plotting table.
Ground control, image orientation and visual interpretation remained with the photogrammetrist. Roads, rivers, buildings, ridges and terrain still had to be identified by an operator. The measured result, however, could move directly into a digital production chain instead of first becoming linework on a cartographic manuscript.
By 1988, seven analogue stereoplotters had been fitted with GeoVision equipment to support digital mapping. Contemporary reporting describes their use for structured urban land bases, local-authority and utility work, forestry datasets, line-of-sight profiles and other products intended for CAD or GIS use. Existing analogue equipment was therefore being adapted rather than discarded. Digital coordinate capture was added to a production environment that still depended on established instruments and operator skills.
The arrangement produced a hybrid workplace. Optical and mechanical equipment remained in use while computers handled increasing amounts of geometry, storage and output. A measurement extracted from a stereo model could be delivered as a coordinate or structured digital feature rather than as a line that another person later had to digitise from paper. This reduced one conversion step between aerial survey and downstream computer systems.
Several separate systems
MAPPAK, Ministry of Works route analysis, computer contouring and digital photogrammetric production developed for different operational reasons. Lands and Survey needed to convert and maintain mapping. Engineers needed to compare alignments and terrain options. Agricultural engineers needed terrain representation from survey observations. Photogrammetric organisations needed efficient ways to convert image measurements into mapping and, increasingly, digital data.
Other digital-geographic systems were developing at the same time. DSIR researchers were processing Landsat imagery as raster data from the mid-1970s, while land-resource teams were digitising mapped environmental units and associated descriptions into the NZLRI and LADEDA environment. Those systems belong principally in the next two chapters. Their chronology overlaps MAPPAK and the photogrammetric work rather than following them in a simple sequence.
The terminology remained mixed. Automated cartography, computer mapping, land information, remote sensing, digital mapping, graphics systems and geographic information systems were not interchangeable labels. A system could use geographic coordinates without being a GIS, and a mapping program could produce digital output without maintaining a reusable spatial database. During the 1980s these previously separate technical traditions increasingly used compatible computers, digital data and commercial software.
Paper remained part of the production system throughout this period. Source maps were digitised, computer-held geometry was plotted back to paper, and analogue photogrammetric instruments worked beside newer analytical equipment. Digital storage changed what could be retained and reused between jobs, but it did not produce an immediate paperless mapping office. Organisations added digital components where they solved particular production problems and retained older equipment where it still worked.
By the end of the 1980s, New Zealand mapping organisations routinely produced and maintained new forms of digital geographic information. Cadastral lines could be stored and edited as coordinates, terrain could be generated through software, orthophotos depended on computer-assisted geometric correction, and photogrammetric measurements could feed directly into CAD and GIS environments. Commercial GIS products entered this environment during the same decade, alongside locally developed and specialist systems. Raster processing and land-resource databases were developing in parallel.
Chapter source notes
1. M. S. Colley, A computer system for a cadastral mapping application, University of Canterbury MSc thesis, 1979, is the principal MAPPAK source. The project has verified Figure 2-1 on printed p.18/PDF p.22 and the interactive digitising figures on printed p.65/PDF p.69 for visual production. The thesis supports an operational computer-assisted cadastral mapping system, not a national-first GIS claim.
2. NZMS 306, edition 1, 1979, provides a period object for the metric conversion problem between 1:63,360 and 1:50,000 national mapping. The National Library record and LINZ provenance are fixed in the visual register.
3. New Zealand Aerial Mapping and analytical-photogrammetry evidence, including the Wild Aviolyt BC1 lineage, supports the transition from optical-mechanical stereoplotting towards computer-controlled photogrammetric production. Equipment dates follow the verified chapter sources.
4. The chapter's source-and-claims appendix remains the working authority for machine-assisted cartography, metric conversion and terrain-computation claims. Chapter 4 takes over where digital imagery and raster processing become the main subject.