| Summary: | Changes in ocean circulation in response to anthropogenic climate change affect ocean biology on a global scale. Based on a previously published empirical model that links ocean circulation to chlorophyll and chlorophyll to primary production, I predict an increase in primary production of 10–27% at the end of the 23rd century under four times pre-industrial atmospheric CO 2. The uncertainty in this prediction largely stems from the reliance on chlorophyll as the only model constraint. Chlorophyll concentrations are difficult to interpret, as they depend on phytoplankton biomass and cellular pigmentation, which adjusts to growth conditions. The objective of this thesis is to bridge the gap between laboratory-based knowledge of physiological adjustments to growth conditions and global satellite observations to reduce ambiguities in the interpretation of chlorophyll concentrations on a global scale. Satellite estimates of phytoplankton carbon and the chlorophyll to carbon ratio (Chl:C), a measure of pigmentation, are the foundation of this work. My main contribution is a re-evaluation of chlorophyll variability in the eastern subarctic Pacific, which updates the old paradigm for seasonal phytoplankton dynamics in this iron-limited region. In contrast to previous studies, I conclude that the consistently low chlorophyll concentrations are caused by a suppression of Chl:C by iron stress, rather than by reduced accumulation of phytoplankton biomass. Field observations during iron enrichment experiments and model simulations confirm that the satellite-observed suppression of Chl:C is consistent with physiological adjustments to low iron. On a global scale, I analyze how phytoplankton biomass and pigmentation interact to yield the spatial structure in surface chlorophyll and I employ a mechanistic photoacclimation model to diagnose the contributions of light, nutrients and temperature to the spatial structure in Chl:C. I further argue that the temporal variability of phytoplankton biomass reflects differences in ecosystem structure and I use an objective clustering algorithm to delineate two large-scale ecological regimes. The spatial extent of the two regimes suggests that the seasonal cycle of light is the main determinant of ecosystem structure. In a second step, I use the same clustering technique to subdivide the regimes into two biomes each, which capture within-regime differences in nutrient availability.
|
|---|