License terms and conditions apply
The Land Classification is a classification of Great Britain into sets of environmental strata, termed land classes, to be used as a basis for ecological survey, originally developed by the Institute of Terrestrial Ecology (ITE) in the late 1970s. The strata were created from the multivariate analysis of 75 environmental variables, including climatic data, topographic data, human geographical features and geology data. The Land Classification has provided a stratification for successive ecological surveys (the Countryside Survey of Great Britain), the results of which have characterised the classes in terms of botanical, zoological and landscape features. Additionally, the Land Classification can be used to stratify a wide range of ecological and biogeographical surveys to improve the efficiency of collection, analysis and presentation of information derived from a sample. There are three versions of the Classification available. 1. The 1990 version containing 32 classes, which was the first to classify all 240,000km squares in Great Britain (building on an earlier 1978 version based on a sample of squares only). 2. A 1998 version, when the Land Classification was adjusted to 40 classes as a consequence of the need to provide National Estimates (from Countryside Survey 2000) for habitats for Scotland in addition to GB as a whole. 3. In 2007, the Land Classification was adjusted once again, to 45 classes, as a consequence of the need to provide Wales-only estimates in addition to those for Scotland and GB as a whole (from Countryside Survey 2007).
Publication date: 2011-10-01
The ITE Land Classification was primarily developed as a sampling strategy for the first Countryside Survey of Great Britain in 1978, carried out by the Institute of Terrestrial Ecology. This initial classification was relatively basic, and created at a time when the idea of widespread ecological survey of areas as big as Great Britain was barely conceivable. However, in the early 1970s the concept of ecological sampling at a smaller scale (e.g. individual woodlands), was commonly accepted and it was R.G.H. Bunce and M.W. Shaw who developed local regional survey into a national ecological sampling system. The important part of the approach was that the sampled areas should be representative of the whole region and, to ensure this, they employed a stratification system. As a result of the earlier work in Cumbria and elsewhere, it was thought that the sample unit of a 1km square was appropriate, being small enough to survey in a reasonable period of time and yet large enough to contain sufficient environmental features to allow differentiation of squares. In the mid-1970s, when this work was being done, computing power was such that software packages like Indicator Species Analysis (ISA) could only operate on a limited number of datasets. Thus it was not possible to create a classification of all 1 km squares in GB. Environmental, physiographic and other mapped data were collected for 1228 squares based on a grid, and the dataset was analysed using ISA to produce 32 classes. Four squares surrounding each of the classified squares were also allocated to classes using the key provided by ISA; thus a total of 6040 squares were classified. The 32 classes were then described based on the average values of the environmental characteristics that were used to generate the classes (for example, average altitude, slope and rainfall, and host of other environmental values). In the 1990 Countryside Survey, it became apparent that for estimations at the regional and local level, the land classification had to be extended so that every square in GB was classified. Although computing power had increased considerably since the first classification, it was still an insurmountable task to collect data for every square in GB at the same level of detail as had been done in the original work. Instead, the major climatic, geological and physiographic factors (or valid surrogates for these) were obtained for each square and then a variety of multivariate classification procedures was tested on the resulting dataset. During the planning stages of Countryside Survey 2000, the need to produce separate reliable estimates of surveyed features for Scotland and England with Wales became a requirement, as did the need to provide statistically reliable estimates of upland habitats in England and Wales. Prior to the 2007 Countryside Survey, it became apparent that in addition to the requirement for Scotland-only results (provided for in the 2000 Land Classification) further adjustments were required to accommodate Wales-only results. Further information is provided in the supporting documents available in the Online Resources.