Fig 1.
Map of the Island of Sardinia showing the Marine Protected Area and corresponding study site.
Tavolara Punta Coda Cavallo Marine Protected Area (Sardinia Island; 40°52’22” N, 9°44’27” E). Three landscapes are mapped (L1, L2 and L3), four habitat classes represent the levels of seagrass fragmentation in each cell. Reprinted from the biocenosis map of the archives of the MPA under a CC BY license, with permission from the managing Director of Tavolara Capo Coda Cavallo AMP, original copyright 2015.
Table 1.
Landscape attributes for each level of fragmentation.
Fig 2.
Spatial distribution of predictor variables.
Spatial distribution of predictor variables within each landscape (L1, L2 and L3): structural attributes (shoot density, canopy height, rugosity and unburied mat,) and prey size.
Fig 3.
Survival curves among landscape areas and habitat classes.
Significant differences were observed for urchin survival among (a) landscape areas (L1, L2 and L3) and (b) habitat classes over 20 days. Levels of significance are represented in Table 2.
Fig 4.
Survival curves among habitat classes within each landscape area.
Significant differences were observed for urchin survival among habitat classes within each landscape area during 20 days. Levels of significance are represented in Table 3.
Table 2.
Coxph-test for predation differences among landscapes and habitat classes.
Table 3.
Coxph-test for predation differences among habitat classes within each landscape area.
Table 4.
Coxph-test used to evaluate the contribution of the covariates (structural attributes, landscape attributes and prey size) to the sea urchin predation risk distribution.
Fig 5.
Spatial distributions of final predation rate and type of predator attacks.
Spatial distribution within each landscape area (L1, L2 and L3) of the final predation rate and type of predator attack suffered by sea urchins. Also, “na” means no available data (dead prey due to stress, see M&M).