Environmental estimates from approximately 276,000,000 grid points were firstly clustered into groups or environments using a non-hierarchical clustering strategy. The relationships between the environments were then defined using the average environmental conditions for each environment. The similarities between environments were explicitly quantified by a dendrogram.

The suite of environmental variables used to define the classification are:

• mean annual temperature
• mean daily minimum temperature of the coldest month
• mean annual solar radiation
• mean winter minimum solar radiation
• mean annual water deficit
• mean monthly ratio of rainfall to potential evaporation
• mean Oct air saturation deficit (temperature adjusted humidity)
• slope- derived from a 25 m digital elevation model
• drainage derived from NZLRI
• acid soluble phosphate (long-term availability of phosphorus)
• exchangeable calcium· parent material particle size
• parent material hardness
• surface age (to identify recent soils)
• chemical limitations to plant growth.

These variables were chosen because of their demonstrated functional linkages to plant physiological processes and their correlations with the spatial distribution of New Zealand's indigenous tree species. The focus on tree orientated variables came about because there is: better data about tree distribution than for other taxonomic groups; and because historically, trees were the dominant feature of New Zealand prehuman landscapes. Land environments can be defined at different scales (see scale of operation.

Both the quantitative approach of the land environments of New Zealand and the qualitative approach of, for example, ecological regions and districts seek to identify ecologically similar areas. The land environments of New Zealand approach aims to group points that are environmentally similar regardless of their location. This gives relatively homogeneous units at detailed scales of mapping. It also means that a unit can be widely dispersed through the landscape. Conversely the ecological regions and districts approach is constrained by the need to group proximate areas into a single spatial unit. This means that a unit defined using the ecological regions adn district approach may have greater environmental variation contained within it

• 20 group classification-primarily paper delivery with electronic
• 100 group classification-primarily paper delivery with electronic
• 200 group classification-primarily electronic delivery with paper
• 500 group classification-primarily electronic delivery

• Level 1-20 groups- 1:2,500,000
• Level II-100 groups- 1:1,000,000 (estimated scale)
• Level III-200 groups- 1:250,000 (estimated scale)
• Level IV-500 groups- 1:50,000 (estimated scale)

• The variables used to compile the classification are treated equally and are used together.
• The classification can be re-worked to define groups at a wide range of scales.

• Climate surfaces developed from metrological station data were used to estimate the climate variables
• New Zealand Land Resources Inventory Database was used to develop the distribution of soils. They were then grouped according to their parent material. Attribute data for each group was then derived from the National Soils Database.
• Slope was derived from a variety of databases over the development process. Most recently a 25 m Digital Elevation Model was used.
• Drainage was derived from the extended soils legend in the Land Resources Inventory Database.

• the representativeness of protected areas
• the representativeness of natural areas remaining
• biodiversity condition
• landscape fragmentation.