A receiver on the trig
At Taiaroa Head in 1994, a GPS survey for a wildlife-management project began with one receiver staying put. It was set over a known trig station on the summit of the headland and left to track the satellites. A second survey-quality receiver was moved to points whose positions had to be established, occupying each for a few minutes. The observations were then processed to determine the vectors from the known station to the new points. Those points became reference sites for later real-time kinematic work and, ultimately, for a spatial information system being developed around the northern royal albatross colony.
The work was part of research led through the rather than a national survey programme. , and described the GPS acquisition at the 1994 Spatial Information Research Centre colloquium, and McLennan’s later thesis preserved more of the method. Trimble Navigation New Zealand supplied receivers and peripherals. carried out photogrammetric interpretation and produced maps used in the spatial database, while graduate students assisted with GPS post-processing and staff supported the field work. The project joined surveying, wildlife management, photogrammetry and GIS in a way that would become increasingly ordinary over the next decade.
The Taiaroa Head team used a fast-static method. The fixed receiver and the moving receiver observed the same satellite signals, including code and carrier-phase measurements on the L1 and L2 frequencies. Specialist processing resolved the relative positions. A coordinate reached the database through observation, processing, checking and data-entry workflows. Getting it involved equipment, a known reference point, field procedure, data transfer and processing expertise.
That combination changed an old part of mapping work. For much of New Zealand’s survey and GIS history, the position of a new feature was derived through an existing framework: a survey mark, a traverse, an aerial photograph, a cadastral plan, a topographic sheet, an address or a known point on a map. GPS introduced another route. Increasingly, the person standing beside the thing being mapped could measure its position directly from satellite signals and then carry that coordinate into the database.
Before GPS
New Zealand surveyors had used earlier satellite-positioning methods before GPS. In 1976 of DSIR’s Geophysics Division published results from Fiordland using an ITT Model 6000 single-channel satellite navigation receiver. This was the older United States Navy Navigation Satellite System, usually known as Transit, not GPS. Davey observed satellite passes at 13 anchorages and compared the resulting positions with coordinates on the New Zealand Geodetic Datum 1949. The exercise found a systematic difference in latitude of about 170 metres between the satellite reference system and NZGD49 in the area studied.
The result formed part of New Zealand’s early satellite positioning work. New Zealand practitioners were experimenting with positions derived from satellites a decade before the first GPS survey currently identified by the historical record. The technologies, constellations and techniques were different, but the underlying problem was already visible: a position calculated in a global satellite reference system did not necessarily sit neatly on the coordinates used by an older national mapping system.
Te Ara identifies a Fiordland survey in 1986 as New Zealand’s first survey using GPS.
By 1989, GPS survey work was more extensively documented. A photograph taken by Evening Post photographer shows a Department of Survey and Land Information surveyor checking GPS equipment on a Wellington hilltop. The receiver equipment occupies a substantial case, with the antenna mounted separately on a tripod. It looks nothing like something that would disappear into a jacket pocket. The photograph leaves the receiver model unidentified and comes from a different event than the 1986 Fiordland survey, but it captures the physical character of late-1980s GPS well: specialised field equipment operated by people who knew what they were doing.
Measuring instead of deriving
A GPS receiver calculates position from timed signals transmitted by satellites. For an ordinary three-dimensional position it also has to solve the receiver’s clock error, which is why several satellites must be visible at the same time. That basic explanation can make the process sound more automatic than early professional practice actually was. Satellite geometry, obstructions, atmospheric effects, reflected signals, receiver quality and the positioning method all affected the result. Ridges, bush and buildings could obstruct satellite signals and affect positioning.
Surveyors could obtain much higher precision than an ordinary stand-alone navigation receiver because they were often interested in the relative position between two receivers rather than accepting each receiver’s independent position. If both observed the same satellites at the same time, many shared errors could be reduced when the measurements were processed together. Carrier-phase observations provided another large improvement because the radio carrier could be measured far more precisely than the coarse ranging code, provided the system could resolve the number of carrier cycles between receiver and satellite.
Static surveying therefore involved occupying points and accumulating observations. Fast or rapid static techniques shortened occupation times when equipment, baselines and satellite conditions allowed it. Differential GPS used a receiver at a known location to provide corrections to another receiver. Real-time kinematic positioning pushed the carrier-phase approach into the field by sending correction information to a rover quickly enough to calculate a high-precision position while the user was still standing there. These were related techniques, but they were not interchangeable names for “accurate GPS”.
The field workflow mattered as much as the receiver. Antenna heights had to be measured correctly. Reference stations had to be known. Batteries, cables, storage and observation schedules had to work. Data then had to be copied, processed and checked. A survey could fail because the mathematics was wrong, but it could also fail because someone entered the wrong antenna height or occupied the wrong mark. Satellites removed some old lines of sight between ground stations; they did not remove field discipline.
Accuracy had a price
By the 1990s GPS could therefore mean several very different things. A geodetic survey might use dual-frequency survey receivers and carrier-phase processing to measure baselines at centimetre or better precision. A mapping project might be satisfied with sub-metre differential positions. A navigation receiver could place a user within tens of metres. All three could be described casually as GPS, which made the acronym rather more precise than the conversation around it.
Correction methods differed in coverage, equipment requirements and cost. ’s 1999 paper on precision farming in New Zealand described sub-metre differential receivers as capable of meeting the needs of precision agriculture, but the correction signal was a significant part of the cost. A user could establish a local base station, use a commercial UHF correction service where coverage existed, or buy corrections delivered through satellite services. Yule described Satlink coverage in Canterbury at NZ$3,600 a year and satellite-delivered correction services through Omnistar and Racal at about NZ$4,000 a year. His worked example allowed NZ$8,000 for the DGPS receiver itself.
Those figures place the technology in its late-1990s setting. Higher accuracy required suitable equipment and survey methods. It could mean a second receiver, a radio link, a correction subscription, suitable coverage and software able to make use of the results. The same paper linked DGPS to yield mapping, GIS, variable-rate application and machinery guidance. Position allowed observations and actions to be attached to the relevant part of the paddock. The coordinate was becoming part of an agricultural information system rather than an end in itself.
Yule also described the practical consequences of guidance. In trials with a fertiliser spreader, ordinary judgement left substantial variation in the vehicle’s path, while DGPS guidance markedly improved the proportion of measured positions close to the intended track. The exact tolerances required by farming differed from cadastral or geodetic surveying, but that was the point. Once receivers could deliver the accuracy needed for a particular job at a tolerable cost, GPS could leave the specialist survey network and become part of operational work.
Corrections are infrastructure
Differential positioning also changed the shape of the system around the receiver. A rover could not improve itself merely by knowing that a more accurate position was desirable. It needed observations or corrections from a reference point whose position was already known. Those corrections had to reach the user, either after the field work during post-processing or in real time through radio or another communications service. The accuracy chain therefore extended beyond the object in the worker’s hand.
New Zealand’s terrain made the correction path a practical design choice. A local base station could give an organisation control over its own reference observations, but it required another receiver and communications equipment. A UHF service could provide real-time corrections without each user operating a base, but radio coverage depended on transmitter locations and terrain. Satellite-delivered corrections reached much larger areas but carried a subscription cost. In Yule’s 1999 agricultural account, commercial UHF coverage was still limited in New Zealand, while the Optus satellite could deliver correction services from Omnistar and Racal.
The economics are revealing. In Yule’s worked precision-farming example, the annual differential correction fee represented about 40 per cent of the operating cost of the extra positioning and control system. A farmer who needed sub-metre guidance was therefore buying access to an information service as well as hardware. The receiver might belong to the farm, but part of its useful accuracy arrived from somewhere else. That model would become increasingly familiar in later high-accuracy positioning, where reference stations, communications networks and correction services sit behind what can appear to the user as one instantaneous coordinate.
Real-time kinematic positioning pushed the same dependency further. RTK could give a rover centimetre-class relative positions while work was underway, but only when carrier-phase ambiguities were resolved and suitable corrections reached the rover reliably. Early radio links imposed range and terrain limits. A lost correction link could turn a high-precision workflow back into waiting, reinitialising or collecting data for later processing. Improved communications allowed positioning corrections to reach receivers away from the office.
This also explains why different sectors adopted different techniques at different times. A national geodetic campaign could justify specialist equipment, planned occupations and extensive processing. A farmer needed a system that worked while machinery was moving. A utility field crew often needed a useful location and the correct asset record more than survey precision. An ecologist might accept post-processing if it produced repeatable monitoring points. The technology spread fastest when the positioning method fitted the work rather than when the most accurate available method was imposed on every task.
The signal gets better
One obstacle to ordinary civilian GPS during the 1990s came from the system itself. The United States deliberately degraded the public signal through Selective Availability. The effect applied globally. It did not prevent precise professional surveying because relative, carrier-phase and differential methods could remove or reduce much of the error, but it made a cheap stand-alone receiver far less accurate than later users would consider normal.
The United States switched Selective Availability off on 1 May 2000. Contemporary official material predicted an improvement in basic civilian positioning from roughly a 100-metre specification to around 20 metres, with real performance often better. Test data published after the change showed an immediate large reduction in positional scatter. Existing receivers gained the improvement without needing to be replaced.
Differential positioning remained essential for surveyors, precision farmers, hydrographers and other users needing metre, sub-metre or centimetre results. The floor shifted instead. An inexpensive receiver that had previously been most useful for broad navigation could now place itself closely enough for many mapping, asset and field tasks. That widened the set of jobs for which a GPS coordinate was worth collecting.
It also changed expectations. A decade earlier, a survey-quality GPS setup had been conspicuous equipment on a hilltop. After 2000, a much smaller unit could provide a field worker with a useful stand-alone position as part of an ordinary work process. The transition from survey instrument to general positioning component was underway.
The coordinate joins the record
GPS became far more valuable to GIS when a position could be attached directly to the record being collected. A point alone says little. A field worker might need to record that the point is a transformer, weed infestation, bird nest, culvert, sampling site or damaged asset, along with condition, date, identifier and notes. The useful change was the connection between measured position and structured information.
The Taiaroa Head work was an early example. Survey control supported a spatial information system in which decades of breeding observations could be related to nest sites and environmental information. The wider system combined GPS observations with databases and field procedures. It supplied spatial control that made other records more reliable and analytically useful. This was a recurring pattern as GPS entered environmental and operational GIS.
By 2004 a and Solid Energy project at Wangaloa Coal Mine in South Otago was using differential GPS with ArcGIS to support rehabilitation monitoring involving vegetation, soils and invertebrate observations. By the later 2000s, coastal research combined RTK GPS with multibeam bathymetry, ArcGIS, ArcSDE and SQL Server. The positioning technology had become one component in larger data-management and analysis workflows.
These systems still depended on quality control after the coordinate was measured. A perfectly good coordinate could be assigned to the wrong asset identifier. A point could be captured beside a pole when the organisation’s data model expected the pole centre, or at a rural gate when another system expected a dwelling. A receiver could be working properly under trees while delivering a worse position than the user assumed. GPS reduced one class of location uncertainty while leaving plenty of ordinary database mistakes available for anyone who missed them.
GIS goes into the field
used mobile GIS to take network information into the field. Development began in August 2001. By April 2002 the electricity network company was introducing 40 Compaq iPAQ handheld computers and six rugged Panasonic Toughbook notebooks for field crews. The project was reported as costing about NZ$200,000, with expected annual savings of about NZ$500,000 from reducing duplicated data entry, errors and trips back to the office for plans.
The equipment is now a small museum of early mobile computing. The Toughbooks ran the GIS used in ’s office. The iPAQs had 64 MB of memory, a 256 MB storage card and Esri ArcPad supplied by . Each PDA also had a Navman GPS device which the contemporary report said could automatically locate the worker to within about 10 metres. developed its own translation between ArcPad and so that field and corporate data could be exchanged.
Connectivity limited the field system. In 2002 had considered wireless access but decided the available bandwidth was not adequate for the amount of mapping data involved. The iPAQs therefore worked largely as synchronised field devices. Staff took data with them, collected or updated information in the field, then plugged the units into an office cradle to exchange data with the corporate GIS. The satellite part was already mobile. Crews synchronised their data when a connection was available.
By August 2003 that boundary had shifted. field staff were using Bluetooth phones and Vodafone GPRS to reach corporate network information from the field. They could search by address or pole number and trace underground cable feeds without returning to the office for a printed plan. GPS, portable GIS and mobile communications were converging, but each had arrived through a different technical route.
’s April 2002 deployment provides a firmly dated example of mobile GIS being introduced for working crews. Earlier handheld and GPS experiments almost certainly existed, and material from the same period describes ArcPad, GPS and Windows CE as an implementation pathway. Operational rollout to working crews marked a stronger milestone than an isolated experiment.
Different jobs, different tolerances
As GPS spread, accuracy became a requirement to be chosen rather than a single measure of technological progress. A geodetic survey establishing national control needed results that could support many other surveys. Engineering and cadastral work had stringent tolerances and formal quality requirements. A utility crew trying to find a buried cable needed enough accuracy to get to the right place, but its field GIS also depended on network diagrams, asset identifiers and local observations. A wildlife team recording a nest or sampling site might accept a different uncertainty again.
Agriculture made the trade-offs particularly visible. Yield mapping and variable-rate application did not require the same accuracy as a cadastral boundary, but errors could still translate directly into wasted fertiliser, poor guidance or misleading maps. A farmer could pay for sub-metre corrections because there was an operational return from placing machinery or measurements more consistently. The value of the GPS system was therefore tied to the economics of the task, not to the smallest number the receiver brochure could print.
Environmental work had similar choices. A repeated monitoring point may need enough positional consistency to ensure a team returns to the same patch over several seasons. A broad species observation may be useful at lower accuracy. An engineering survey around a mine or coast may require differential or RTK methods. GIS made these datasets easy to display together, which could disguise the fact that they had been collected under different positional standards.
Good metadata and field practice became increasingly important for that reason. Users needed to know what coordinate system was being used, whether positions were corrected, how the point had been located and what feature the coordinate represented. A map could draw points with exquisite graphical precision even when the field observations behind them were much rougher. Digital display never had much difficulty looking more certain than the evidence.
People between antenna and database
The spread of GPS also created work that sat between surveying and GIS. Someone had to plan observations, set up bases, manage receiver files, apply corrections, transform coordinates and decide whether a result was good enough for the task. Equipment suppliers trained users and helped organisations choose between navigation, mapping and survey systems. GIS staff designed feature classes and attribute forms so a field position arrived with enough information to be useful. Database staff then had to reconcile new observations with records that might have been created years earlier from plans or manually digitised maps.
The Taiaroa Head project makes that mixture unusually visible. Surveying expertise established the spatial control. Photogrammetry supplied mapped detail. Wildlife staff supplied the biological observations and management context. Post-processing converted raw receiver observations into usable coordinates. The spatial information system brought those parts together. The project depended on surveyors, field crews, database staff and GIS specialists as well as GPS equipment.
The same was true in operational organisations. ’s field crews carried the devices, but the system depended on office GIS, data, ArcPad translation, asset-management rules, hardware support and later mobile communications. Precision agriculture combined farmers and machinery operators with positioning services, controllers, GIS software and agronomic interpretation. As satellite positioning became easier to use, some of the specialist work became less visible rather than disappearing. The field user increasingly saw a coordinate, while reference networks, software, data models and quality checks did more of their work in the background.
The maintained positioning framework also depended on named specialists. LINZ records Paula Gentle working with the GeoNet and PositioNZ GNSS networks and the International GNSS Service, Dionne Hansen as a senior geodetic/geophysical analyst, and Nic Donnelly working on the national reference frame and analysing land movement after the Canterbury and Kaikōura earthquakes. Their work sat between satellite signals, survey practice and the coordinates used by other spatial systems.
From GPS to GNSS
For the early part of this history, GPS is the correct term because the United States system is what the New Zealand sources describe. As receivers later became capable of using signals from several satellite constellations, GNSS became the more accurate umbrella term. The vocabulary changed gradually, and historical writing needs to follow the technology rather than replace every old reference to GPS with a modern label.
More satellites and more signals improved availability and resilience, especially where parts of the sky were obstructed. Receiver electronics also became smaller, cheaper and less power-hungry. Correction delivery moved from local radio links and dedicated services towards continuous reference networks and internet distribution. These developments made high-accuracy positioning easier to use in ordinary work, although the underlying distinctions between autonomous, differential and carrier-phase methods remained.
The field device changed with it. Separate receiver, data logger and office-processing stages increasingly gave way to integrated controllers and mobile GIS applications. Features and attributes could be captured together. Corrections could arrive while the rover was working. Positions could be checked against existing layers immediately rather than after a day’s field work. None of these changes happened on one date, but together they shortened the distance between observing something and seeing it correctly placed in an organisational database.
Position and interpretation
Chapter 17 followed the construction of national road, address and property-location systems. GPS did not make those systems redundant. A receiver can tell a user where its antenna is in a coordinate framework. It does not know that the place is 24 Example Road, that the pole beside it is asset P10327, that the land is part of a particular parcel, or that the location lies within a statistical area. Those meanings come from other records and institutions.
Mobile field GIS combined measured positions with existing asset and observation records. ’s crews did not use GPS as a substitute for the electricity-network database. They used position alongside asset maps, pole numbers, addresses and cable connectivity. At Taiaroa Head, GPS did not replace decades of albatross breeding observations. It helped give those observations reliable spatial relationships. In precision farming, DGPS did not decide where fertiliser was needed. It allowed a machine and its data to act on a location chosen through other measurements and analysis.
Direct coordinate capture still required mapping judgement. The antenna might be at the gate while the feature being described was the house. A user might stand safely beside a road rather than on the centreline being updated. A utility asset could be underground and inferred from other evidence. GIS still needed a data model that stated what the coordinate represented.
A global coordinate meets a moving country
GPS also exposed a deeper problem. The satellites operated in a global geocentric reference environment, while much New Zealand mapping and surveying still depended on NZGD49 and coordinate systems built for an earlier era. A GPS receiver could produce an internally good position that did not overlay older national data correctly if the datum relationship was ignored. The issue had been foreshadowed by the Transit observations in Fiordland in the 1970s, and it became harder to avoid as GPS use expanded.
During the 1990s repeated GPS campaigns became part of the work used to establish a new national geodetic framework. Five repeat surveys between 1992 and 1998 supplied observations for the initial realisation that became NZGD2000. The response went well beyond buying better receivers. New Zealand needed a coordinate framework compatible with satellite positioning and capable of dealing with a country whose crust does not stay put.
Emergency dispatch made the consequences of digital location data unusually immediate. In 1996 Police replaced 29 district control rooms with three communications centres in Auckland, Wellington and Christchurch using the Communications and Resource Deployment system, CARD. Police later recorded that the first job entered into the new system was logged at 14:36 on 28 October 1996 and concerned youths removing guttering from a scout hall in Hamilton. The event was unremarkable, which is useful historically. A national digital dispatch system began its recorded working life with an ordinary address that someone still had to find.
A 2005 independent review showed how dependent the centralised model had become on mapping. Call takers were relying increasingly on digital mapping tools, while the review still encountered complaints about lost local knowledge and recommended better aerial mapping, rural rapid-number data and GPS coordinates for rural properties. It also identified a much smaller problem with larger consequences: the Intergraph CARD Address Search field could delay confirmation of a caller's street if the search was not broadened correctly. Location technology had made national dispatch possible, but it had not abolished the basic question at the start of an emergency call: where, exactly, is this person?
By the end of the period covered here, the practical change was already established. A coordinate no longer had to arrive mainly through an existing map or a chain of ground measurements carried back to the office. It could increasingly be measured where the work happened, attached to a feature, checked against other data and transferred into GIS. The next problem was ensuring that millions of those measured positions meant the same thing once they reached the national map.
Positioning becomes school material
Positioning technology also became part of public geospatial education. LINZ’s 2016 LEARNZ field trip, Where are we?, focused on navigation and positioning on sea, land and air. The 2018 Land, Sky and Space trip extended that story through SBAS, aircraft navigation, the Awarua satellite ground station, autonomous vehicles, forestry, drones and modern surveying. In 2020 Get Outdoors Safely returned to the practical end of the same technology, using maps, GPS, distress beacons and search-and-rescue operations to show why accurate location information affects both everyday navigation and emergency response.
Sources · 1
Chapter source notes
1. Transit precursor. F. J. Davey's 1976 DSIR Geophysics Division work in Fiordland documents satellite positioning with the US Navy Navigation Satellite System/Transit and the documented comparison with NZGD49. This prevents the GPS narrative being presented as the beginning of all satellite positioning.
2. 1986 GPS claim. Te Ara identifies a 1986 Fiordland survey as New Zealand's first GPS survey. The original survey record remains unrecovered, so publication wording must attribute the claim or describe 1986 as the earliest identified GPS survey rather than a proven absolute first.
3. 1989 DOSLI photograph. Jon Hargest's Evening Post photograph held by the Alexander Turnbull Library shows the specialist physical character of late-1980s GPS equipment. The receiver model is unverified.
4. Taiaroa Head. The 1994 SIRC paper by Bruce McLennan, Albert Chong and Martin Purvis and McLennan's later thesis documents the fast-static GPS workflow, known trig station, moving receiver, L1/L2 code and carrier-phase observations, post-processing and later RTK-related work. Retain Peter Hunt and Department of Conservation support only in the roles supported by those sources.
5. Precision farming. Ian Yule's 1999 New Zealand Grassland Association paper documents New Zealand DGPS use, correction-service options and period costs, including Satlink at NZ$3,600 per year, satellite corrections at about NZ$4,000 per year and NZ$8,000 for the receiver in the worked example.
6. Selective Availability. Official GPS.gov/United States government material documents removal of Selective Availability on 1 May 2000 and the improvement in ordinary civilian autonomous positioning. Professional differential and carrier-phase surveying before that date was not restricted to about 100 metres.
7. Vector mobile GIS. The 15 April 2002 New Zealand Herald report and the documented August 2003 follow-up document development from August 2001, deployment of 40 Compaq iPAQs and six Panasonic Toughbooks, ArcPad/Smallworld translation, Navman GPS, initial cradle synchronisation and later Bluetooth/GPRS connectivity. Retain Vector as the earliest firmly dated sustained production-mobile GIS deployment in the project corpus, not as a proven national first.