Figure legends
Figure 1. Two simplified examples of the effect of turf species on canopy-forming macroalgae. A) In this example, the canopy kelp is negatively affected by two species that live in the understory below. This competitive effect can be measured experimentally, through the removal of the turf and crust understory. If the abundance (or biomass, etc.) of the canopy increases in the absence of the turf and crust species, the interaction is estimated as competitive. B) However, if these species facilitate the canopy kelp, then the removal of turfs and algal crusts should have a negative effect on the canopy.
Figure 2. Locations of the studies included in the meta-analysis, by study type. On the right is the distribution of studies in the Northeast Pacific Ocean, where a high density of studies have been conducted.
Figure 3. The overall effect of different turf functional groups on the canopy (kelps and Fucales combined). Each point in grey represents an individual measured interaction, with means and 95% confidence intervals in black. If log response ratio (LNRR) is negative, then the effect of turf species on the canopy is facilitative. If LNRR is positive, then the effect is competitive.
Figure 4. The experimental effect of different turf functional groups on the canopy, by depth. Each point is a raw data point, scaled in size by the study sample size, with 95% confidence intervals shaded around each line. Note that the slope and intercept of the coralline crust and non-coralline turf lines do not statistically differ from zero.
Figure 5. Across latitude (top), mixed evidence for increasingly facilitative interactions at higher latitudes. Below, patterns across life history stage (raw data in grey, means and 95% confidence intervals for the mean in black). For life history stage, observations were collected primarily on the recruit canopy stage (see Results: Observational studies). Plot across latitude includes only coralline turfs, plot across life history includes all canopy and turf taxa.