Summary
This multi-regional study assessed whether statistical models predicting wild bee abundance in oilseed rape fields could be reliably transferred across different regions and years. Whilst mass-flowering crop cover consistently reduced pollinator density across all study regions, the effect of semi-natural habitat was variable between regions, and overall model transferability was limited. The findings highlight important constraints in using landscape-level land-use variables to extrapolate pollinator responses across geographical and temporal scales, likely because such variables inadequately capture local variation in pollinator resources.
Regional applicability
The study included United Kingdom data from the broader five-region assessment (Sweden, Germany, Netherlands and UK), making findings directly relevant to UK agricultural pollination contexts. However, the conclusion that landscape models have limited transferability suggests that UK-specific parameterisation would be necessary for reliable local predictions, rather than relying on models developed in other regions.
Key measures
Bumblebee and solitary bee abundance; landscape-scale cover of mass-flowering crops, semi-natural habitats, grasslands and forests; model transferability assessed via cross-validation across spatial and temporal scales
Outcomes reported
The study evaluated the spatial and temporal transferability of statistical models linking wild bee abundance to landscape-scale land use in winter oilseed rape fields across five regions in four countries over three years. Models showed limited transferability across regions and time, with consistent negative effects of mass-flowering crop cover but variable responses to semi-natural habitat cover between regions.
Topic tags
Dig deeper with Pulse AI.
Ask about this record, its theme or its relevance to UK farming and policy. Pulse AI uses selected catalogue evidence and cites the sources it draws on.