Understanding scales of density-dependence to improve the use of resources in benthic mussel aquaculture

Shellfish aquaculture is considered a sustainable way to help meet rising protein demands worldwide. In shallow coastal dynamic ecosystems mussels can be cultivated directly on the seabed, however this method has a low return as mussels exposed to natural environments risk dislodgment, high predation rates, sedimentation and competition. The formation of spatial patterns in natural mussel beds, that result in ‘organized patchiness’, is thought to be an adaptive mechanism to reduce population losses. The driver and effects of this patterning need to be disentangled at multiple spatial scales in which patterns are observed. With a field experiment we aimed to understand how small-scale density (actual cover) and patterning (perimeter: area ratio of clumps and number of mussel layers) can be altered by manipulating large scale density (re-laying biomass), that farmers could control during seeding activity. Within this study we considered the interplay between environmental conditions (manipulating flow rate with the use of large mesh cages) and density for pattern development and persistence, and the repercussions of this on mussel productivity (growth and condition). We further investigated local scale processes, such as the role of within-clump biological activity (biodeposition), that may be a predictor for the larger scale observations of losses and persistence relative to density. We found that manipulating density by controlling seeding biomass from boats is not an accurate predictor of actual seabed density and resulting patterning. The growth and condition of the mussels was only influenced by local scale effects, resulting in high ‘within clumps’ variation. Aiming for an intermediate density to avoid both excessive fragmentation and excessive layering may be viewed as an optimal strategy to maximise returns, but we encourage the incorporation of the hierarchy of multiple scales of density in future studies of patterning that will allow the inclusion of these effects in a model of growth and productivity