Estimating scale-dependency in disturbance impacts: El Niños and giant kelp forests in the northeast Pacific.

Recent discussions on scaling issues in ecology have emphasized that processes acting at a wide range of spatial and temporal scales influence ecosystems and thus there is no appropriate single scale at which ecological processes should be studied. This may be particularly true for environmental disturbances (e.g. El Niño) that occur over large geographic areas and encompass a wide range of scales relevant to ecosystem function. However, it may be possible to identify the scale(s) at which ecosystems are most strongly impacted by disturbances, and thus provide a measure by which their impacts can be most clearly described, by assessing scale-dependent changes in the patterns of variability in species abundance and distribution. This, in turn, may yield significant insight into the relative importance of the various forcing factors responsible for generating these impacts. The 1997-98 El Niño was one of the strongest El Niños ever recorded. I examined how this event impacted giant kelp populations in the northeast Pacific Ocean at 90 sites ranging from central Baja California, Mexico to central California, USA. These sites spanned the geographic range of giant kelp in the Northeast Pacific and were surveyed just before, immediately following, several months after, more than 1 year after, and nearly 2 years after the El Niño. I used a hierarchical sample design to compare these impacts at five spatial scales spanning six orders of magnitude, from a few meters to more than 1,000 km. Variance Components Analyses revealed that the El Niño shifted control over giant kelp abundance from factors acting at the scale of a few meters (local control) to factors operating over hundreds to thousands of kilometers (regional control). Moreover, El Niño resulted in the near-complete loss of all giant kelp throughout one-half of the species' range in the northeast Pacific Ocean. Giant kelp recovery following El Niño was far more complex and variable at multiple spatial scales, presumably driven by numerous factors acting at those scales. Recovery returned local control of giant kelp populations within 6 months in southern California, and within 2 years in Baja California.