Recent increases in the availability of time-series satellite data allow regional variability in producer biomass and productivity to be evaluated on seasonal to decadal scales.  These data are especially valuable for ecosystems where appropriate field surveys are time and labor intensive.  Here, we combined satellite imagery with diver sampling to assess local and regional changes in the biomass of giant kelp (Macrocystis pyrifera) at unprecedented temporal resolution.  The LANDSAT 5 sensor provided us with coverage of the entire Santa Barbara Channel approximately every two months from 1984-present.  We paired these data with monthly diver surveys conducted by the Santa Barbara Coastal Long Term Ecological Research Project from 2002-present.  Our objectives were to: 1) develop new methods for estimating giant kelp biomass from an extended time series (>25 years) of satellite imagery and 2) assess changes in kelp forest biomass across multiple temporal and spatial scales.

Results/Conclusions

We successfully developed an automated classification technique where each pixel in an image was modeled as a combination of kelp and water spectra thus producing continuous maps of fractional kelp coverage.  Monthly diver observations of canopy biomass in fixed plots at two kelp forest sites were well correlated with satellite determinations of fractional kelp coverage (r² = 0.63) allowing us to examine the dynamics of giant kelp biomass across multiple spatial scales.  The relationship between diver measured plot scale (~40 m) and remotely assessed site scale (~1 km) estimates of biomass varied between sites and depended on the relative location of the plot and the size of the kelp forest at each site.  Correlative analyses involving a range of oceanographic and climatic variables (e.g., swell height, sea surface temperature, nutrients, ENSO and PDO indicies) provided insight into the potential drivers of local and regional dynamics in giant kelp biomass and possible responses of the giant kelp system to changes in these drivers.  Linking remotely acquired data with long-term ecological field measurements can facilitate a better understanding of the patterns and drivers of biomass and primary production for a multitude of terrestrial and aquatic ecosystems.  Synthesizing long term observations at multiple scales is vital to understanding and predicting ecosystem responses to a changing climate.