Comparing sources of variability in agronomic weed seed predation: Time, space, and habitat

Background/Question/Methods

Post-dispersal seed predation can reduce arable weed seedbanks, providing a desirable ecosystem service to farmers. Seed predation can be substantial, but the magnitude of seed predation is highly variable. To help farmers maximize benefits received from seed predators, we must understand sources driving this variability. 

Other studies show that time, space, and habitat contribute to variability in the magnitude of seed predation. To compare the relative importance of these sources of variability, we conducted a farmscape-level series of spatial-explicit seed predation assays across diverse crop and non-crop habitats on a typical Maine diversified organic vegetable farm. Seed predation, invertebrate activity-density, and habitat attributes were measured at 215 sites in mid-October 2011, and August 2012; and at a subset of 132 sites in September 2012 and October 2012.      

Results/Conclusions

We observed significant inter- and intra-annual variation in seed predation. Total and invertebrate seed predation was greater in October 2011 (22.7% total; 5.76% invertebrate) than October 2012 (13.8%; 2.34%). Total and invertebrate seed predation decreased from 40.5% and 18.4% (respectively) in August 2012 to 13.8% and 2.34% in October 2012. 

Correlogram analysis indicates that seed predation was randomly distributed across the landscape. This contradicts our hypothesis, and the results of other studies; we expected seed predation to be spatially aggregated. It is possible that habitat heterogeneity “masked” the effect of space in our study system. 

We found significant habitat type and vegetative cover effects. Seed predation was greater in crop and riparian forest than in softwood forest, meadow, or mowed grass habitats. Within the crop habitat only, seed predation was positively correlated with leaf area index and other measures of vegetative cover. Across habitat types, however, vegetative cover did not support higher seed predation rates. The effects of habitat type and vegetative cover varied significantly across time. 

Our results suggest that time and habitat factors are much more important regulators of seed predation than space in our study system.