Britain Finally Unites 277,000 Vegetation Plots in One National Database

Britain’s landscapes have been studied, mapped, and argued over for more than a century, yet until now the country has never maintained a single centralised repository for its vegetation plot data. That gap has finally been closed. A new database report published in the journal Vegetation Classification and Survey introduces the National Vegetation Plot Database for Great Britain and the Crown Dependencies, known by the acronym GBNVPD and registered under the Global Index of Vegetation Plot Databases with the code EU-GB-007. Its first version is not a modest pilot: it contains 277,070 samples drawn from 200,733 plots, documenting 4,463,300 occurrences of 4,086 accepted taxa, gathered from surveys conducted between 1949 and 2024.

The scale of the achievement becomes clearer when you consider where the data came from. Version 1.0 of the GBNVPD incorporates 63 individual datasets, of which 52 are newly submitted to the European Vegetation Archive and 9 to ReSurveyEurope, a database of resurveyed plots that underpins analyses of long-term vegetation change across the continent. Until this submission, only four datasets from Great Britain were present in the EVA, representing 82,565 samples, or 29.8 percent of what the new national database now holds. The GBNVPD therefore makes available 194,519 samples from 52 datasets to the EVA and 97,099 samples from 9 datasets to ReSurveyEurope for the first time, plugging one of the largest holes in the European vegetation data landscape.

The road to this point was long and littered with near misses. Discussions in 1981 between John Rodwell’s National Vegetation Classification team and the Nature Conservancy Council led to a database held by the Unit of Vegetation Science at Lancaster University, built from roughly 13,000 plots surveyed during the NVC project and expanded to 31,450 plots with contributions from collaborators across the country. It was registered as the first British database in GIVD, under code EU-GB-001. But as Rodwell himself later observed, aftercare of the NVC project deliverables was limited, and except for particular groups of vegetation types, new data were never formally quality-controlled and incorporated into any kind of UK national vegetation database.

Several subsequent attempts to build a national repository also fell short. A register of permanent vegetation plots developed by Hill and Radford in the 1980s was never collated and standardised. Mark Hill’s BIGSHOT project, an ambitious plan for a biological inventory of georeferenced species and habitat occurrences over time, failed to secure funding but did rescue a small quantity of unstandardised datasets. Ian Strachan’s effort to create a national vegetation plot database for Scotland, commissioned by Scottish Natural Heritage, came tantalisingly close to completion, with all plot data standardised for entry into Turboveg, but was never published. Only recently has a resurgence in vegetation plot data rescue efforts at the United Kingdom Centre for Ecology and Hydrology provided the firm foundation needed to realise a centralised national database.

The technical standardisation behind the GBNVPD is where much of the scientific rigour lies. All taxon names in the original datasets were aligned to version 20241114a of the United Kingdom Species Inventory, a comprehensive taxonomic backbone that allowed the original taxon concept names to be preserved where valid, satisfying the critical requirement for unambiguous recording of different taxon views. A lookup table connects every taxon in the database to the latest accepted UKSI names, and a further lookup links accepted taxa to EuroSL, the Euro+Mediterranean synonymised species list, where equivalent taxon concepts exist. This dual alignment means British plot data can now be compared directly with continental datasets without the taxonomic friction that has historically hampered such analyses.

Vegetation classification was handled with equal care. Each plot in the database was matched against three systems: the EuroVegChecklist syntaxa using the R implementation of the EUNIS expert system, Great Britain broad habitats using a count of indicator species, and National Vegetation Classification communities using pseudo-quadrat methodology implemented in the R package RMAVIS. Five cover-abundance scales were present in the constituent datasets, including DAFOR, Domin, percentage cover, rooted frequency, and nested presences. All original cover values were retained as per EVA convention, while DAFOR and Domin values were converted to percentage cover to enable cross-dataset comparison. The database itself is elegantly simple, consisting of ten tables that store species presence and cover, plot attributes, the top-five fitting NVC communities, the top-five fitting broad habitats, the top-fitting EuroVegChecklist syntaxa, taxonomic lookups, dataset metadata, and source information. The whole database is versioned and stored as a DuckDB database held internally by UKCEH.

Inclusion criteria were deliberately broad to rescue as much data as possible. Only three requirements were applied: confidence in the reliability of the survey methodology, confidence in the reliability of vascular plant identification, and absence of experimental treatments. Plots of any size were accepted, with any number of attributes, and even with absent or incomplete coverage of bryophyte and lichen taxa. The geographic scope encompasses Great Britain and the Crown Dependencies of the Isle of Man, the Bailiwick of Guernsey, and the Bailiwick of Jersey, but deliberately excludes Northern Ireland and the Republic of Ireland, which are covered by the Irish Vegetation Database. In a handful of cases, control plots from experimental datasets were included where they sampled phytosociological units otherwise poorly represented, such as the rare Arenarion norvegicae communities of mountain ledges.

The content of the database reveals both the strengths and biases of British vegetation surveying. The plots were surveyed between 1949 and 2024, with approximately 40,000 to 60,000 plots recorded in each of the 1970s, 1990s, 2000s, and 2010s, and a noticeable lull in the 1980s. The samples are distributed across the country at a mean of 10 samples per square kilometre, though with a pronounced bias towards the Sheffield region, which was heavily sampled during the Unit of Comparative Plant Ecology surveys. Based on top-fitting NVC matches, the database encompasses all habitats as defined by the NVC, with mesotrophic grasslands dominating at 29.7 percent of samples, followed by woodlands at 12.7 percent, mires at 11.1 percent, open vegetation at 9.7 percent, and calcifugous grasslands and montane communities at 8.3 percent. Aquatic habitats appear under-represented with only 700 samples as the top match, though this is certainly an underestimate given that 1,139 samples were used to define the original NVC aquatic communities.

The species composition of the database is highly skewed, reflecting the prevalence of lowland semi-improved and improved grassland in national monitoring schemes. Only 78 taxa account for 50 percent of all occurrences, and more than 50 percent of taxa occur fewer than 50 times. The most frequent species are ubiquitous grassland plants with broad niches: Yorkshire-fog (Holcus lanatus) at 1.9 percent of occurrences, creeping bent (Agrostis stolonifera) at 1.5 percent, sweet vernal-grass (Anthoxanthum odoratum) at 1.5 percent, ribwort plantain (Plantago lanceolana) at 1.4 percent, and white clover (Trifolium repens) at 1.3 percent. Species richness is correspondingly low, with 70.4 percent of plots over one square metre containing only zero to five species per square metre. This distribution highlights the value of targeted survey efforts such as the Botanical Society of Britain and Ireland’s Threatened Plants Project, without which many rare and scarce taxa would be poorly represented.

The future development of the GBNVPD is ambitious. The team plans to integrate new vegetation plot data from ongoing monitoring programmes including the Environment and Rural Affairs Monitoring and Modelling Programme of Wales, the England Ecosystem Survey, the Countryside Survey, the National Plant Monitoring Scheme, and the Environmental Change Network. Historic datasets also need rescuing: the England and Wales sand dune surveys of the 1990s remain undigitised. Plot attributes will be gap-filled, alternative survey methods such as limestone pavement grike surveys will be incorporated, and a more sophisticated database structure will be designed for experimental plot data. The database is available under a discretionary mixed-access regime, with coordinate precision deliberately restricted for sensitive datasets such as the Countryside Survey, where 22.5 percent of samples are limited to 10,000-metre precision to obscure plot locations. What emerges is a living national resource, one that transforms decades of scattered British fieldwork into a coherent, continuously updated scientific instrument capable of feeding directly into European and global analyses of vegetation change.

Subject of Research: A National Vegetation Plot Database for Great Britain

Article Title: A National Vegetation Plot Database for Great Britain

Article References: Marshall, Z., Wood, C., Mountford, E., Rodwell, J., Strachan, I., Pescott, O., Tatarenko, I., Hodgson, J., Walker, K., Latham, J., Maskell, L., Britton, A. J., Pakeman, R. J., Dargie, T., Mitchell, R., Jones, L., Hester, A., Ross, L., Dalrymple, S., … Smart, S. M. (2025). A National Vegetation Plot Database for Great Britain. Vegetation Classification and Survey, 6, 181-190. https://doi.org/10.3897/vcs.160378

Image Credits: AI Generated

DOI: 10.3897/VCS.160378

Keywords: National, Vegetation, Plot, Database, Great, Britain, scientific research