Kelp forest ecosystems dominate 150,000 km of global temperate coastline, rivalling the coastal occurrence of coral reefs. Despite the astounding biological diversity and productive ecological communities associated with kelp forests, patterns of species richness and composition are difficult to monitor and compare. Crustose coralline algae are a critically important substrate for propagule settlement for a range of kelp forest species. Coralline covered cobbles are home to hundreds of species of benthic animals and algae and form a replicable unit for ecological assays. Here, we use DNA metabarcoding of bulk DNA extracts sampled from cobbles to explore patterns of species diversity in kelp forests. The data from 98 cobbles within kelp forest ecosystems at three sites in Central California show the presence of 2,037 Molecular Operational Taxonomic Units (MOTUs) and 79 MOTUs assigned up to the species level with >95% similarity to current databases. We are able to detect spatial patterns of important management targets such as abalone recruits, and localized abundance of sea stars in 2012. Comparison of classic ecological surveys of these sites reveals large differences in species targets for these two approaches. In order to make such comparisons more quantitative, we use Presence/Absence Metabarcoding (PAM), using the fraction of replicate cobbles showing a species as a measure of its local abundance. This approach provides a fast and repeatable survey method that can be applied for biodiversity assessments across systems to shed light on the impact of different ecological disturbances and the role played by marine protected areas.