GIS across government

By the late 1990s, a government task could use GIS without becoming a GIS project. The specialised systems described in Chapter 10 had established that departments could maintain national spatial databases and digital mapping environments. Desktop software had widened access, while census boundaries, cadastral data, addresses, topography and other national datasets were becoming easier to obtain in forms that ordinary analysts could use. Location could now enter a policy paper, operational response or regulatory system without the organisation first building a mapping department around it.

Administrative and analytical work also began using geographic information. Treasury analysts wanted to understand differences between regional economies. Biosecurity staff needed to find infected and potentially exposed apiaries during an incursion. Fisheries managers had to relate species, catch and management areas to the marine environment. Mining and petroleum administrators dealt with legal permit areas that had to be defined, checked and maintained geographically. In each case, the map sat beside administrative records, scientific evidence or policy analysis rather than replacing them.

Central government remained organisationally uneven. Some agencies had specialist GIS teams, some bought prepared datasets and desktop software, and others used spatial functions embedded in larger information systems. Restructuring also moved functions between departments during the 1990s and 2000s. Government therefore used several architectures rather than one common model. The shared pattern was that more public-sector work could assume the existence of digital geographic information and staff who knew how to use it.

In June 2000 the Treasury published a working paper by Suzi Kerr and Jason Timmins titled Economic Geography & Spatial Statistics: Theory and Empirics of New Zealand Regions. It examined regional economic characteristics and explicitly used GIS as part of the analytical method. Critchlow Associates supplied Census96 data and ArcView software. The material included geographic boundaries and demographic, household, labour-force and socioeconomic information organised at meshblock, area-unit, territorial-authority and regional-council levels.

The researchers used existing geographic information for their analysis. Statistics New Zealand had already done that work, and Chapter 16 follows its development. Nor were they implementing a departmental GIS production system. They were policy researchers obtaining prepared spatial data and software, then using them to examine regional differences through maps and spatially organised statistics. The case sits a long way from the mapping agency production environments of the previous decade.

The geographic units gave the researchers several scales at which to examine the same country. Meshblocks provided fine-grained population geography, while area units and local-government boundaries supported broader regional comparison. Economic variables could be mapped rather than read only as columns in national tables. GIS made geographic pattern easier to inspect and allowed analysis to be organised around places rather than one national average, while causal interpretation still depended on the underlying evidence.

The supply arrangement also records how government GIS could depend on organisations outside the department. Treasury received both software and prepared Census96 information from Critchlow rather than assembling every component itself. Statistics New Zealand remained the source of the census and statistical geography, while a commercial supplier packaged the data for desktop use. Policy analysis could therefore draw on a chain of national data production, commercial preparation and desktop software without Treasury becoming a producer of those underlying spatial datasets.

By 2000 this kind of use required far less technical infrastructure than it would have a decade earlier. An analyst could work with regional boundaries and attributes on a desktop computer rather than submit a mapping request to a specialist production unit. The work was still recognisably GIS, but the organisational role had changed. Geography had become another analytical dimension available to a policy team.

The same year produced a much less leisurely government use of location. Varroa destructor was confirmed in South Auckland on 11 April 2000. The parasitic mite attacks honey bees and had not previously been established in New Zealand. The Ministry of Agriculture and Forestry began an exotic-disease response that included movement controls, field inspection and a delimiting survey intended to establish how far the infestation had spread.

The response geography changed as evidence arrived. A controlled area was declared on 15 April and revised later in the month. Inspectors needed to know where registered apiaries were located, which sites had been visited, where positive findings had occurred and which areas remained to be checked. Beehives were movable assets, so the response also involved controlling movement across the boundary being defined. A paper list of beekeepers and a map of apiaries were both useful, but the operational problem required the two to agree. The boundary defined the area used for administration or analysis. It affected where hives, equipment and people could move while the response team was still learning where the infestation actually was.

Government approved another $1.35 million on 1 May to extend the survey. By 28 May contemporary response reporting recorded visits to 2,120 apiaries containing 40,461 hives. The same report described work under way to improve the mapping. Apiary maps were being standardised on the 1:250,000 terrain map, and detailed digitising had begun so the maps could be used with GIS. GIS work developed as the response progressed.

That sequence is more useful than claiming that GIS ran the Varroa operation from the first day. The initial response relied on the information and maps available when the incursion was discovered. As the survey expanded, the limitations of inconsistent or less readily usable apiary mapping became more costly. Standardising the base and digitising the apiary information allowed the expanding field record to be handled in a common spatial framework.

The response team still needed beekeeping records, laboratory confirmation and people visiting properties. GIS did not find mites by itself. It organised a geographic workload in which thousands of apiaries had to be related to survey status, movement controls and an evolving understanding of the infestation. New findings could alter the area receiving attention, which in turn changed where staff and resources needed to go next.

The maps also had administrative consequences. Movement controls applied to defined areas, while decisions about eradication or management depended partly on the apparent distribution of infected sites. A boundary on the map therefore represented more than a picture of the outbreak. It affected what beekeepers could move and where government response activity was concentrated.

MAF Biosecurity had been established in 1999, so the Varroa response occurred inside a recently reorganised national biosecurity system. Its mapping work illustrates the kind of operational GIS that became more common in government during this period. A programme or incident generated administrative records, field observations and locations, and staff needed them to remain connected while decisions were being made.

Later technical reconstruction by veterinary epidemiologist Robert Sanson shows how far the spatial work developed after those first weeks. Varroa_sim was programmed in ArcView using Avenue and treated 9,229 registered North Island apiaries with valid map-grid coordinates as the population at risk. The stochastic model represented local spread, movement within a beekeeper's operation and the seasonal concentration of hives for pollination, while eradication assumptions could be changed through menus in the map interface. The work also exposed a quieter data problem. Apiary map-sheet references were subsequently converted automatically to full New Zealand Map Grid coordinates and validation was added to stop impossible locations entering the database. GIS was therefore doing two jobs: modelling where infestation might move and forcing the underlying location records to become more dependable.

The Ministry of Fisheries developed another model in the following years. The National Aquatic Biodiversity Information System, usually shortened to NABIS, was developed in 2001 as a geospatial data reporting and management tool. It brought together spatial information relevant to New Zealand fisheries and marine biodiversity, including species distributions, fisheries management areas and commercial catch. By the middle of the decade it was operating as a web-based GIS used in fisheries and biodiversity planning, monitoring and policy work.

Marine administration is inherently geographic. Quota management areas have boundaries. Commercial catch is reported against defined areas and locations. Species distributions vary across depth, habitat and region. Marine reserves, fishing restrictions and other management zones overlap a physical environment that has no roads or parcel boundaries to orient the user. A spatial system provided a common interface for information that otherwise sat in separate research, management and administrative records.

NABIS allowed users to build maps from selected layers rather than receive only a fixed published map. A user could combine a base map with biological distributions, management boundaries and catch information. The system therefore served several types of work from the same maintained information. A policy analyst could examine an area under consideration, a researcher could compare biological and management geography, and an external user could inspect information without obtaining the ministry’s internal GIS software.

The service also became a continuing programme rather than a one-off web publication. A 2009 study of spatial information in the New Zealand economy recorded three full-time-equivalent staff focused on NABIS management and development. A May 2008 survey of about 2,000 users informed work on a second-generation system. Forty survey respondents estimated an average productivity improvement of about nine per cent.

The staffing is as revealing as the interface. Someone had to maintain the layers, check new material, manage the application and decide how the information would be presented. A web GIS did not eliminate specialist work behind the screen. It changed the way that work reached users. The ministry could maintain spatial information centrally while policy staff, researchers and external users interacted with selected parts of it through a common service.

The NABIS lineage also shows how the boundary between GIS and ordinary information systems was becoming less obvious. A user interested in orange roughy catch or a fisheries management area did not need to think first about GIS software. The task was to find and compare fisheries information by location. The mapping technology remained underneath the service, but the business subject increasingly defined the experience.

Minerals and petroleum administration provides a different government case because geography is written directly into the legal object being administered. A permit applies to an area. The boundary therefore has to be described precisely enough for applicants, regulators and other users to identify where the rights apply. By the 2000s the regulatory framework was explicitly describing those areas through maps, coordinates and official reference systems.

The Crown Minerals (Minerals Other than Petroleum) Regulations 2007 set requirements for permit maps and the description of permit areas. Boundaries could be defined using coordinates, land parcels, bearings and distances, geographic features, coastlines and other specified references. The petroleum regulations similarly required mapped permit areas and information about the datum and projection used. The permit was an administrative record whose legal extent depended on spatial definition.

Permit officers could use the system without becoming GIS specialists. The permit system had to preserve geometry accurately enough to support regulatory work. Applications had to be checked against existing rights and the geography of the area concerned. Changes, relinquishments and new permits altered the permit pattern over time. A spatial database was a natural way to maintain that changing set of areas alongside permit identifiers and status information.

Later New Zealand Petroleum and Minerals services make that connection visible. Permit-register information is linked with GIS databases and web maps, and users can view permit areas, print maps, download spatial files and use web services. Current mapping guidance accepts digital geographic formats and specifies the national datum and projection used for submitted data. Those later services should not be backdated into 2007, but they are a clear endpoint for the regulatory lineage already visible in the mapping requirements.

The Crown Minerals case differs from Treasury and Varroa because the geometry forms part of the regulated object itself. It is part of the regulated object. An incorrectly placed permit boundary can affect whether an area is available for application and how rights overlap. Coordinate reference information, source geometry and area definitions therefore become part of administrative quality control.

The system also depends on data maintained elsewhere. Coastlines, cadastral information, coordinate frameworks and place references originate in other national systems. Geological information may inform exploration without defining the legal permit itself. The regulatory database sits among these sources and has to distinguish the permit geometry it controls from reference data supplied by other organisations.

The four cases used different software and had different institutional purposes, but the working pattern is recognisable. Treasury combined spatial boundaries and statistics for policy analysis. MAF connected registered apiaries, field survey and control areas during an operational response. Fisheries maintained a reusable spatial information service around biological and management data. Crown Minerals administration treated mapped areas and coordinates as components of permits.

None required a single central-government GIS. Each agency used national or sector data appropriate to its work. Statistical geography came from Statistics New Zealand. Topographic and coordinate frameworks came through national mapping and geodetic systems. Agencies added their own administrative or scientific information and maintained the parts for which they were responsible. The resulting government spatial environment was distributed long before cloud services made that architecture easier to describe.

GIS teams and technical staff remained part of the arrangement, but their role increasingly involved supporting other users. They prepared datasets, maintained layers, solved coordinate problems, built repeatable processes and connected geographic records with other information systems. Policy staff and programme teams could then use the spatial result without understanding every part of the production chain. The organisational split was similar to other forms of information technology: specialists kept the infrastructure working while a larger group used the service.

This wider use increased the cost of poor maintenance. An old boundary could affect a regulatory decision. A missing apiary could distort an operational picture. A fisheries layer whose date or source was unclear could be reused outside the context in which it was produced. A statistical boundary joined to the wrong census vintage could produce a technically successful but historically mismatched analysis. Metadata and stewardship, discussed in Chapter 21, were practical requirements inside these ordinary workflows.

Government restructuring added another complication. Functions moved between agencies, and names changed. Data, systems and staff had to move with them if continuity was to be preserved. The Ministry of Fisheries functions later entered the Ministry for Primary Industries, while minerals and petroleum administration now sits within the Ministry of Business, Innovation and Employment. A spatial database could outlive the department name on the report that first described it.

By the early 2010s the government GIS user was no longer easy to identify from a job title. Some were dedicated GIS analysts. Others were economists, policy advisers, biosecurity staff, fisheries scientists, permit administrators or programme managers using an application in which geography was already built in. Internal viewers and web systems further reduced the need for every user to have a desktop GIS licence.

The map consequently became less useful as a measure of whether an agency had mature GIS. A policy paper might contain only a few static figures even though its analysis depended on spatial datasets. An operational system might display locations without allowing the user to edit geometry. A regulatory database could use spatial validation behind forms and permit records. GIS capability was spreading into workflows that did not advertise themselves as mapping systems.

Maps remained central. Treasury published them because regional pattern was part of the analysis. MAF produced outbreak and control-area maps because people needed to understand the response geography. NABIS was explicitly map based, and permit administration still required mapped boundaries. The larger change lay behind the map: maintained datasets linked to policy, programme, scientific and regulatory records.

By this stage government agencies could also expect spatial information from one another. A department did not have to create its own national coordinate framework, statistical geography or topographic base before beginning a project. That dependence made national data maintenance part of ordinary government capability. It also meant that decisions made by one data custodian could affect applications in agencies far removed from the original production system.

The early government GIS story was about departments acquiring systems and converting maps into digital databases. By the 2000s the technology had moved deeper into administrative work. Geographic information was being used to examine economic regions, manage a biosecurity response, organise fisheries information and define regulated permit areas. None of those activities existed primarily to make maps.

The systems remained uneven and agency specific. Some were desktop analyses, some operational datasets, some public web services and some regulatory databases. They did not converge on one vendor or one central government platform. They shared the assumption that location could be stored, queried and maintained alongside the other information needed to run the programme.

That assumption changed the place of GIS in government. A new spatial task no longer required an argument that geographic information belonged in the work. The questions were increasingly about data quality, identifiers, access, maintenance and how the spatial record connected with the business system around it. Network operators had already been confronting the same questions with electricity, telecommunications, water and other infrastructure, where connectivity between assets became as central as their location.