Despite the impressive progress that has been made over the last twenty years in the use of satellite data and other remote sensing to gather information about the land surface, there are some things that can only be recorded through the use of technology no more impressive than a map, a clipboard and a pencil. In particular, the detailed characteristics of vegetation and the presence of individual plant species, so meaningful in ecological terms, are best identified through field survey.

A major element of the earlier GB countryside surveys of 1978, 1984 and 1990 was a sample survey of land cover, landscape features and vegetation. By repeating these again in CS2000, an invaluable long-term data series has been established which has no equal in GB or elsewhere.

The aims of the field survey were to:

• estimate the extent and distribution of widespread habitats in GB;

• characterise widespread habitats in terms of their land cover and botanical composition and to assess changes in these characteristics;

• derive indicators of sustainable development for the wider countryside;

• provide accessible databases containing information about the state of the British countryside;

• provide ground reference data for the calibration and validation of Land Cover Map 2000.

To survey land cover in detail, within one field season, is not practicable using a census survey approach - it would take far too long and would be very expensive. To obtain cost-effective but reasonably reliable results, the field survey used a sampling approach; the sampling unit that has been used in all countryside surveys is a 1 km square and mapping is done at a scale of 1:10,000. In CS2000, 569 squares were surveyed. These squares were taken at random from different land types, as defined by the ITE Land Classification, and thereby represent the full range of environmental conditions in GB.

In CS1990, 506 of these squares were surveyed and 61 new squares plus two replacement squares have been included to allow improved estimation of certain habitats (especially in the uplands of England and Wales) and to give a better sampling rate in the individual countries of GB. A separate but compatible survey of Northern Ireland has also been carried out this summer and this allows results to be expressed for the UK as a whole.

An early task was to define what is meant by widespread habitats. Since the publication of results from CS1990, the UK Biodiversity Action Plan has listed both Broad and Key Habitats. Some of these, such as lowland heath and chalk grassland, are relatively scarce and cannot be assessed accurately using the current sampling approach. Most of the Broad Habitats however are reasonably common or well distributed and it is these that are termed ‘widespread habitats’. These habitats form the primary framework for the reporting of stock and change results from CS2000 but to provide continuity, some results are also be expressed using the same classification of land cover as used in earlier surveys.

More than 60 field surveyors were employed in the project. In each square, surveyors working in teams of two mapped the land cover and landscape features (such as hedgerows, walls, woods and ponds). They also recorded all the higher plants and a restricted list of lower plants from 27 vegetation plots, varying in size from 4 m2 to 200 m2. Each square took up to six days to complete and the survey owes a deep debt of gratitude to the farmers and landowners who gave permission for access and who tolerated the surveyors being on their holdings for such a length of time.

The field survey generated millions of data items, many of them in mapped form. These were entered onto computers using Geographical Information Systems and other purpose-built software packages. Despite using an army of data experts, this task, together with the analysis and reporting of the data, took a long time.

Since the start of the work, the project team have looked forward to seeing what changes have taken place between 1990 and 1998. Have the various agri-environment schemes had a significant impact on features, such as hedgerows, in the countryside? Has the previous decline in plant diversity in the lowlands been reversed? What ecological effects have afforestation and building development had on the environmental capital of the landscape? Only surveys of this type can provide the answer to these sorts of questions in the wider British countryside. Accounting for nature: assessing habitats in the UK countryside, published in November 2000, answers these questions and many more.