Résumé
“Buffer capacities” has been defined in ecology as a holistic concept (e.g., Integration of Ecosystem Theories: A Pattern, second ed. Kluwer, Dordrecht, 1997, 388pp), but we show that it can also be worked out in mechanistic studies. Our mechanistic approach highlights that “buffering capacities” can be depleted progressively, and, therefore, we make a distinction between current and potential “buffering capacities”. We have applied this concept to understand the limited “local stability” in seagrass ecosystems and their vulnerability towards structural changes into macro-algal dominated communities. We explored the following processes and studied how they confer buffering capacities to the seagrass ecosystem: (i) net autotrophy is persistent in
Zostera noltii meadows where plant assimilation acts as a sink for nutrients, this contrasted with the
Ulva system that shifted back and forth between net autotrophy and net heterotrophy; (ii) the
Z. noltii ecosystem possesses a certain albeit rather limited capacity to modify the balance between nitrogen fixation and denitrification, i.e., it was found that in situ nitrogen fixation always exceeded denitrification; (iii) the nitrogen demand of organoheterotrophic bacteria in the sediment results in nitrogen retention of N in the sediment and hence a buffer against release of nitrogen compounds from sediments, (iv) habitat diversification in seagrass meadows provides shelter for meiofauna and hence buffering against adverse conditions, (v) sedimentary iron provides a buffer against noxious sulfide (note: bacterial sulfide production is enhanced in anoxic sediment niches by increased organic matter loading). On the other hand, in the coastal system we studied, sedimentary iron appears less important as a redox-coupled buffer system against phosphate loading. This is because most inorganic phosphate is bound to calcium rather than to iron. In addition, our studies have highlighted the importance of plant–microbe interactions in the seagrass meadows.