Forestry GIS

Plantation forests require repeated mapping as stands, roads and management activities change. Stands are planted, thinned and harvested. Roads are built and extended, stream crossings change, firebreaks are maintained, and compartments that once held trees of one age may contain something quite different after the next rotation. Forest maps needed updating as stands, roads and management activities changed.

Foresters had managed this problem long before GIS. Forest maps divided estates into compartments and stands, aerial photographs helped identify boundaries and condition, and field inventory supplied measurements of stocking, diameter, height, species and volume. Separate records carried planting year, silvicultural treatment and expected yield. A manager could combine those sources manually, but every change created more work keeping the map, inventory and planning records aligned.

Computers entered forestry through modelling and inventory before the GIS label became common. A Ministry of Works and Development and New Zealand Forest Service forestry model was documented in 1979–80 as a tool for forecasting future developments in forest industries. The Forest Research Institute also developed a substantial modelling culture around growth, yield and management. These systems dealt with forests numerically, but a model of future timber supply was not the same thing as a geographic information system.

The 1983 Forest Research Institute manual on indigenous forest survey gives a detailed view of computer-supported inventory before GIS became routine. Field measurements were collected through defined survey methods and tied to maps and aerial photographs. Records were coded for computer processing, with punched cards still part of the workflow, and programs were used for checking, summarising, grouping and analysing the observations. The system depended on a chain from field plot to coded record to computer output.

Location remained essential even when much of the processing was statistical. An inventory plot had to be related to the forest area it represented. Aerial photographs and maps provided the geographic framework, while the computer held the measurements and calculations. The two parts could be managed separately, but doing so made later spatial questions cumbersome. Finding every plot associated with a particular area, stand condition or management problem could require matching paper maps with tabular printouts and coded records.

Plantation forestry created an even stronger case for joining the parts. A stand is both a place and a management record. Its boundary defines the area to which planting year, species, stocking, pruning, thinning, condition and yield estimates apply. If the boundary changes but the database does not, area calculations become wrong. If the attributes change but the map does not, the map becomes a confident picture of obsolete information.

Forestry therefore generated a strong use case for digital spatial databases. The practical task was to maintain mapped operational units whose attributes could be updated and used repeatedly in inventory, road planning, harvesting and other work. Map production remained part of the workflow, but the maintained stand and compartment records increasingly sat behind the map.

By 1989 the Ministry of Forestry was describing GIS explicitly as part of New Zealand forestry work. Andrew Dunningham and S. Thompson published “Use of geographic information systems in New Zealand forestry applications” in the Commonwealth Forestry Review. Their paper described the need within Ministry of Forestry service and research work for efficient storage, retrieval, transformation and display of spatial information. The examples used TERRASOFT, a GIS running on a personal computer.

One case dealt with a forest of about 550 hectares. TERRASOFT was used for mapping, inventory and harvesting planning, with additional work on viewshed analysis and visual-impact assessment for landscape design. The system stored more than stand outlines. Mapping, inventory and harvesting records could be used alongside terrain and visibility analysis, allowing the same spatial environment to support both current forest records and planned operations.

The personal-computer platform also differed from the mainframe and specialist workstation environments found elsewhere in early New Zealand GIS. A smaller forestry operation did not necessarily need a national database or a large central computing facility. A PC-based system could hold the spatial information for an estate or project and support local planning. The scale of the system could match the scale of the forest being managed.

TERRASOFT was one of the systems used in New Zealand forestry. Forest inventory databases already existed, New Zealand Aerial Mapping was producing digital mapping, and other GIS systems were operating elsewhere in government and science. The 1989 paper documents GIS already being applied to ordinary forestry tasks in New Zealand without establishing that the TERRASOFT installation was the first.

Forest management works through units that are useful because they can be tied to decisions. A stand boundary might enclose trees planted at roughly the same time and managed under a common regime. Compartment and estate boundaries provide other levels of organisation. Roads, landings, streams and terrain cut across that structure and affect how the trees can actually be reached, managed and harvested.

A GIS could keep those features in the same coordinate framework while retaining different attributes for each. The stand layer could carry planting and treatment information. Roads could be classified by status or capability. Terrain data could support slope or visibility analysis. Harvest plans could select the stands due for treatment and examine the access required to reach them.

Area calculation was an everyday benefit. A stand drawn on paper could be measured manually, but every boundary revision required another calculation. In GIS, editing the polygon changed the area available to later queries and reports. That did not guarantee the boundary was right. It reduced the repeated arithmetic once the geometry had been checked.

The same applied to harvesting. A schedule based only on age and volume could identify which stands were ready, but the operational problem also involved access, slope, neighbouring stands, streams and sometimes landscape effects. Dunningham and Thompson’s viewshed work records an early New Zealand attempt to include some of those surrounding conditions in the same spatial system used for forest management.

Aerial photography remained one of forestry’s main spatial sources during this transition. Plantation boundaries, roads, clearings and differences in stand condition could be interpreted from photographs, while repeated photography provided a record of change. The Forest Research Institute literature of the late 1970s already included work on aerial photography in exotic forest inventory and the use of inventory for plantation planning. GIS added another destination for the interpreted information rather than replacing the photography itself.

New Zealand Aerial Mapping had long supplied forestry and other land-based industries with aerial survey. Its move into analytical and then digital photogrammetric production during the 1980s meant measurements from photographs could increasingly be delivered as digital coordinates or CAD/GIS data. Chapter 3 followed the Wild BC1 and later GeoVision equipment behind that transition. For forestry clients, the practical change was shorter movement from aerial survey to a computer-held forest map.

The source still required interpretation. A photograph could show a visible edge between stands or the line of a road, but a database needed to know what the feature represented. Field information and management records were still required for attributes such as planting year, treatment history, stocking or expected yield. Digital mapping reduced one part of the conversion chain while leaving forestry knowledge at the centre of the record.

The institutional setting changed abruptly in 1987 when the New Zealand Forest Service was disestablished. Plantation and commercial functions were moved into new corporate arrangements while conservation responsibilities were redistributed elsewhere. Staff, paper records, computer files and inventory systems did not all move through that reorganisation in the same way.

A later Westland research project records one particularly stark case. In 1986 an inventory database used by the study had been stored on a New Zealand Forest Service computer. After the Forest Service was dissolved, the computer files could no longer be found. The surviving material consisted of one private copy of a database printout and one map showing the inventory plot locations.

The later researchers re-entered the inventory into GIS so the records could again be linked to spatial survey information. The case should not be generalised into a claim that Forest Service databases were broadly lost during restructuring. It does document how a digital dataset could disappear while paper derived from it survived. A computer file was only durable if somebody retained the system, media, documentation and responsibility needed to preserve it.

The Westland case also records a preservation failure. Paper could be awkward to query and expensive to copy, but it could sit on a shelf through an organisational change. A database tied to a particular computer and institutional owner could vanish when that environment disappeared. Later GIS allowed the Westland inventory to be reconstructed because the printout and map had survived long enough to be entered again.

By the middle of the 1990s desktop GIS was being used directly in Ministry of Forestry operational programmes. An August 1995 New Zealand Forestry article described an East Coast erosion-control forestry grants project based in Gisborne. The programme had begun the previous year and was already monitoring more than 16,000 hectares. MapInfo was used to produce maps, record land use and calculate the hectares associated with grants.

This was different from the 1989 TERRASOFT research and demonstration cases in scale and administrative setting. The GIS supported a live programme that needed to know where afforestation was occurring and how much land qualified for grant calculations. Mapping and measurement were part of the same administrative workflow. The contemporary report also identified Critchlow Associates as the New Zealand MapInfo master agent, placing commercial desktop GIS support beside government forestry use.

The East Coast programme connected forestry with erosion control, so it also crossed the boundary between production and environmental management. Trees were being established partly as a response to unstable land rather than solely to create a timber crop. GIS could hold the planting areas, land-use information and programme boundaries needed to manage that work. The forestry application was therefore operational without fitting neatly into one sectoral box.

By 1995 a programme office could use a commercial desktop package for mapping and calculations without maintaining its own specialist GIS development environment. The software still required trained staff and reliable data, but its technical setting differed from the earlier workstation systems used by large mapping agencies and research units. Commercial support and training could also be obtained locally.

Forestry retained specialist models alongside GIS rather than replacing them. Systems such as STANDPAK were developed to model radiata pine growth, management and utilisation. They could compare silvicultural regimes, rotation age, site quality, stocking and expected financial outcomes. Their strength lay in modelling what might happen within a stand, while GIS supplied the geography of where the stands were and how they related to terrain and neighbouring features.

By the later 1990s researchers were linking these forms of information more directly. New Zealand work on forest visualisation used stand boundaries created in TERRASOFT and ARC/INFO, digital elevation data and outputs from STANDPAK to create scenario-based views of future forest structure. The stand geometry could be draped over terrain while modelled age or management states changed through time. This allowed researchers to examine how planned operations might alter the appearance and structure of a forest landscape before the changes occurred.

The workflow was still technically awkward. Different software expected different formats, coordinate systems and metadata, and the 1997 visualisation work records manual editing and conversion needed to move data between systems. Integrating the systems required file conversion and checks. Much of the useful work sat in the conversion steps between specialised tools.

The resulting simulations extended forestry GIS beyond inventory and map production. They could support landscape planning, visual assessment and comparison of management scenarios. A stand was no longer represented only by its current mapped boundary and attributes. It could also be linked to a model of what that stand might look like after harvesting, replanting or several years of growth.

Forestry exposed the maintenance problem more quickly than many other spatial datasets. A geological boundary might remain serviceable for decades. A plantation map could be wrong within a season if harvesting, road construction or replanting had not been entered. Stand attributes also aged continuously as trees grew and treatments occurred.

A useful forest GIS therefore required routines for updating both geometry and attributes. New aerial photography could reveal changed boundaries and new roads. Field crews and inventory teams supplied measurements that altered stand records. Harvesting created a new state immediately, and re-establishment started another management cycle. Staff continued to maintain the database after the initial conversion.

Errors could also move in both directions between map and management. An incorrect stand boundary affected area and volume estimates. An outdated planting year could place a stand in the wrong harvest schedule. A road shown as available when it was not could distort access planning. The value of the GIS depended less on the sophistication of the display than on whether the underlying records were current enough for the decision being made.

A forest company or agency might manage a smaller geographic area than a national topographic or land-resource database while revising its records much more frequently. Harvesting, road construction, silvicultural treatment and replanting continually altered the spatial and descriptive record. Forest GIS therefore operated as an updating system as well as a mapping system.

Interpine’s technical articles make routine forestry GIS work visible. David Herries’ 2014 GeoMaster and ArcMap guide describes mapping unstocked gaps from imagery or GPS, updating the stand record and preserving the reason for the change.

Source WEB-INTERPINE-PEOPLE-2026 · Interpine technical guide, 2014

A 2001 company account traces the business to Mike King’s forestry contracting and names Donald King as general manager.

Source WEB-INTERPINE-ORIGIN-2026 · Interpine company history, 2001

By the end of the 1990s GIS sat more naturally beside the other computer tools used in New Zealand forestry. Forest researchers were combining stand maps with digital terrain, growth models and landscape visualisation. Operational programmes were using desktop GIS for mapping and area calculation. Aerial-survey providers could deliver digital mapping instead of forcing every client to digitise paper output again.

Forestry organisations adopted several system architectures. Different organisations used different GIS packages, modelling systems, inventory databases and data sources. Plantation owners also differed in estate size and business requirements. Spatial data became part of the maintained forest record.

The 1989 TERRASOFT case provides a clear early benchmark because it joins mapping, inventory and harvesting planning inside a named GIS. The East Coast MapInfo project shows desktop GIS embedded in an operational forestry programme several years later. The Westland inventory case records the opposite problem: spatial information can be lost when institutions and systems change, even when the original data collection involved substantial public effort.

By 1993 the Forest Research Institute was using GIS in a much more explicit analytical workflow at Kaingaroa. Barbara Höck and Tim Payn, working with J. W. Shirley of Forestry Corporation, had 1,643 forest compartment boundaries in digital form and imported them into ARC/INFO. The compartment identifiers linked 744 known site-index values to the spatial data, while the GIS supplied compartment-centre coordinates to a geostatistical analysis and received the estimated values back for mapping. The job shows a step beyond computerised forest inventory: the forest's management units had become spatial objects that could connect inventory records, statistical estimation and map production.

A 1998 paper by Forest Research scientists Barbara Höck, Tim Payn and Andrew Dunningham shows what this looked like after the institute became a Crown Research Institute. Their work combined remote sensing, permanent sample plots, foliage measurements, soil-map units, GIS and geostatistics. One trial treated satellite pixels as 25-metre ground cells and tested them against measured forest data; another loaded growth and foliage plots into GIS and overlaid them with soil units while the team worked on a better link between MARVL and GIS. Forestry scientists used GIS to compare and analyse their information. GIS allowed staff to compare field inventory, imagery and environmental information.

Forestry therefore entered GIS through work it was already doing. Inventory, aerial photography, road planning, growth modelling and harvesting all existed beforehand. GIS gave those activities a common geographic framework and made the stand or compartment a maintained digital object rather than only a shape on a paper map. Science more broadly developed the same pattern, with geological, ecological, hydrological and other research groups building their own spatial databases and analytical systems around different kinds of evidence.