The repeatability of vegetation classification and mapping.

Area of Science:

• Ecology
• Conservation Biology
• Botany

Background:

• Habitat mapping using plant communities is crucial for conservation planning and monitoring.
• Existing mapping protocols often suffer from subjectivity and observer bias, raising concerns about reliability.

Purpose of the Study:

• To assess the consistency of habitat maps generated using the National Vegetation Classification (NVC), a widely adopted system in the UK.
• To identify sources of variation in mapping and evaluate the reliability of NVC for conservation site description and monitoring.

Main Methods:

• Seven surveyors independently mapped the same upland site using NVC classification within a five-week period.
• Spatial correspondence between the generated maps was quantitatively assessed at different hierarchical levels of the NVC.

Main Results:

• Spatial agreement between maps decreased significantly with lower levels of NVC sub-classification.
• Average agreement was 77.6% at the habitat level, dropping to 34.2% at the community level and 18.5% at the sub-community level.
• Discrepancies in classification, particularly confusing similar species compositions, were the primary source of variation, not boundary placement.

Conclusions:

• The study raises serious concerns regarding the current use of the NVC for ecological site description and monitoring due to low inter-surveyor consistency.
• Surveyor experience appeared to influence agreement, but effort-related factors did not explain map discrepancies.
• Further research is recommended to determine the sources of variation and improve the consistency of vegetation mapping.