Community-scale seedbank response to less intense rotation and reduced herbicide input at three sites

Species diversity and abundance in weed seedbanks were measured at the beginning, and after 3 and 6 yr of the TALISMAN experiment at ADAS Boxworth, Drayton and High Mowthorpe. Fifty species were distinguished in the seedbanks of treatments comprising winter and spring rotations and full and reduced herbicide input. After 6 yr, the number of seedbank species in the main treatments ranged from 11 in the most intense to 26 in the least intense management. Principal co-ordinate analysis showed that the seedbank community diverged over time between sites and additionally between the two rotation treatments at Boxworth. Only four species were common to all sites but community-scale descriptors indicated that the change in seedbanks following reduced inputs had common features at the three sites. First, species-area relations showed a greater relative increase in number of taxa at a plot scale than in the site as a whole, indicating lower inputs encouraged the common weeds more than rare species. Second, both the mean and the standard deviation of the species-abundance distribution at 6 yr increased in response to reduced inputs. As a result, the number of species slightly more than doubled as herbicide units were reduced from four to one while the total number of seeds increased by two orders of magnitude. Some causes of seedbank change were identified: spring-germinating species tended to increase when spring cultivation was common in the rotation, while some competitive weeds had a relatively low abundance when they were targeted by herbicide. The results are discussed in relation to optimising the various functions of the seedbank.