Stability and change in kelp forest habitats at San Nicolas Island

Kelp forest communities are highly variable over space and time. Despite this complexity it has been suggested that kelp forest communities can be classified into one of 2 states: kelp dominated or sea urchin dominated. It has been further hypothesized that these represent “alternate stable states” because a site can remain in either of these states for decades before some perturbation causes a rapid shift to the other state. Our research group has maintained a subtidal community monitoring program for 38 years at San Nicolas Island consisting of twice-annual scuba-based surveys at 6 sites distributed within 4 regions around the island. Three types of perturbations are thought to be relevant to subtidal community dynamics at San Nicolas: (1) physical disturbances in the form of major storm and El Nino/ Southern Oscillation (ENSO) events; (2) invertebrate diseases, which periodically decimate urchin populations; and (3) the reintroduction and subsequent increase of sea otters (Enhydra lutris nereis). These 3 perturbations differ in spatial and temporal specificity; physical disturbances and disease outbreaks occur periodically and could affect all 4 regions, while sea otter predation has been concentrated primarily at the West End sites over the last 15 years. The different types of perturbations and the duration of the time series at the kelp forests at San Nicolas make the data set ideal for testing the “alternate stable state” hypothesis. We use nonmetric multidimensional scaling (NMDS) to examine spatial and temporal patterns of community similarity at the 4 regions. In particular, we evaluate support for the existence of stable states, which are represented on NMDS plots as distinct spatial clusters. Community dynamics at each site approximated a biased random walk in NMDS space, with one or more basins of attraction and occasional jumps between basins. We found evidence for alternative stable states at some sites, and we show that transitions from one stable state to another may be influenced by interactions between multiple perturbations.