A functional assessment of the response of grassland vegetation to reduced grazing and abandonment

Question: Predicting the impact of land‐use change on vegetation is vital to understanding how biodiversity and ecosystem function may respond. Is it correct to assume that abandonment is an extreme form of grazing reduction? Location: Borders and central Scotland. Methods: The analysis used data sets from two identical experiments where the impacts of two unfertilized, extensively grazed treatments and one unfertilized abandoned treatment were compared against the species dynamics of a pasture subject to normal, productive grazing management over a 16‐year period. Initial multivariate analysis using Principal Response Curves was used to assess if particular traits were associated with either extensive or abandoned treatments, and was checked using univariate tests of individual traits. RLQ analysis followed by clustering into response groups was used to assess if species behaved in a similar manner between sites. Results: For many traits/attributes the shift in value or proportion was approximately linear across the extensification treatments as grazing was removed. However, certain traits showed step changes and quadratic responses. Leaf dry matter content, an important effect trait, was in the latter group. Most traits/attributes and species behaved similarly at the two sites. However, traits such as regenerative strategy, seed length, longevity and mass and seed bank type behaved differently, indicating that they are not predictable response traits. Conclusion: The results indicate that responses to grazing removal during extensification are largely straightforward and largely independent of species pool. However, there are discrepancies that suggest that simple analyses of the impacts of land‐use changes such as grazing reduction may hide more complex responses.