Activities  Environment Modelling grid Swiss Environmental Domains

Swiss Environmental Domains (SwissED) is an environmental classification of key climatic, geologic and topographic variables influencing both natural and anthropogenic processes at various scales. It represents a new spatial framework to analyse data about our environment (e.g. biodiversity, land cover, demography, agriculture, economical activities) that is not replacing existing ones but simply complementing them.

Although Environmental Domains were initially developed as a tool for biodiversity management, it has a much wider application. Actually, environmental factors that control the distributions of many land-based plants and animals (temperature, water supply, etc.) are also factors that provide major constraints on human land uses such as agriculture, horticulture and forestry.

Compared to a traditional ecosystem classification, generally depending on subjective synthesis of multiple information sources, this multivariate classification approach presents several advantages because it is adjustable, categorical, repeatable and scalable.

The method was originally developed in New Zealand by John Leathwick at Landcare Research organisation (Leathwick J.R. et al. 2003, Conservation Biology, 17: 1612-1623.).

This GRID-Europe project, mandated by the Swiss Federal Office for the Environment (FOEN), started in February 2007.

Methodology (clustering)

SwissED was inspired from several similar initiatives developed in Australia, New Zealand, USA and Europe.

It follows a quantitative and reproducible approach composed of two

i) first a non-hierarchical classification to group a sample of pixels representing Switzerland into a 120 domains,

ii) second a hierarchical classification of these 120 domains into 100, 50, 25 or 10 domains. These domains can be coloured following the result of a PCA
analysis where red corresponds to a gradient of temperature, green a gradient of calcareous content and blue a topography gradient.

The first 10 domains were named according to their environmental characteristics: calcareous reliefs, molassic flats and hills, quaternary hills and valleys, crystalline slopes , dry quaternary flats, calcareous midslopes, calcareous upper slopes, crystalline crests, crystalline quaternary slopes and calcareous crests.

Figure 1: Swiss Environemental Domains at Level I (10 groups), Level II (25gr.), Level II (50gr.) and Level IV (100gr.)


The project team believes that SwissED will bring a new and necessary spatial framework to underpin environmental research and management in Switzerland at a range of scales.

Future applications could be:

• providing a framework for reporting on the state of the environment;

• identifying the most efficient use of limited financial resources for biodiversity conservation;

• management, including management of protected natural areas and other areas of land with
high biodiversity values;

• identifying sites where similar problems are likely to arise in response to human activities, or
where similar management activities are likely to have a particular effect;

• identifying the geographic extent over which results from site-specific studies can be reliably
extended; and

• designing stratified sampling strategies.

Figure 2: Percentage of vineyards per 10 swiss domains (upper left), 10 biogeographic regions (upper right), 25 domains (lower left) and 26 states (lower right)


•The strong climatic, geologic and topographic gradients found in Switzerland represent the ideal pre-conditions for building environmental domains;

•When compared to traditional spatial frameworks, the maps produced when representing statistics (e.g. land cover) on SwissED return more realistic spatial patterns and surface areas;

•SwissED does not replace previous spatial framework but can bring a valuable complementary tool to represent environmental data;

•SwissED are in line with similar developments made across the world at continental, regional or national levels; and

•SwissED were developed for general purposes analyses without trying to weight the input variables, they could therefore be improved by targeting a specific need (e.g. biodiversity, land cover, agriculture).