We characterized the ecological status of each island by quantifying the abundance of kelps, understory algae, coralline algae, etc. at randomly selected sites, using the same methods that have been employed by us and others over the past 30 years (Estes et al. 2010). Originally, we identified potential study sites by laying a grid over a map of each island, marking every place a grid line intersected the coast; these marks were later assigned GPS waypoints. During the 2014 research cruise, we randomly selected and resampled six sites per island, or in the Semichi Islands (Alaid, Nizki, and Shemya)—island group—as this level of sampling is sufficient to determine the ecological status of an island (Estes et al. 2010). We performed identical community surveys at the sites we studied with respect to algal reef bioerosion (see associated metadata forms and datasets).

At each site, one diver sampled twenty 0.25-m^2 quadrats along the 20 foot depth contour, taking a random number of kicks between quadrats. Patches of unconsolidated substrate were rare on the rocky reef; however, if one was encountered while sampling, the diver took an additional random number of kicks and again deployed the quadrat. In each quadrat, s/he counted the stipe density of all kelps (by species) and estimated the abundance (percent cover) of fleshy red algae, sessile invertebrates, encrusting coralline algae (almost exclusively Clathromorphum nereostratum), branching coralline algae, and other algae such as Desmarestia or Codium species. Percent cover was visually estimated on a scale of 1-6, where 1 = 0-5%, 2 = 6-25%, 3 = 26-50%, 4 = 51-75%, 5 = 76-95%, and 6 = 96-100 % cover.