Abstract :
[en] In the framework of an ongoing EU project (PUMPSEA: “Peri-urban mangrove forests as filters and potential phythoremediators of domestic sewage in East Africa”), we compared biogeochemical characteristics of the water column in two constrasting Tanzanian mangrove systems during the dry season: (i) Mtoni, located within the perimeter of Dar es Salaam and which is subsequently under significant anthropogenic influence, and (ii) Ras Dege, a pristine site located ~30 km south of Dar es Salaam. Despite inherent differences in the environmental settings and hydrology of both systems, we wanted to evaluate whether the inputs of sewage in the impacted site would be detectable using a range of biogeochemical parameters, including physico-chemical and nutrient data, dissolved methane, concentrations and stable isotope signatures of different C and N pools (dissolved inorganic carbon, dissolved and particulate organic carbon, particulate nitrogen). Overall, our data suggest that for most of the parameters examined, the range of values and their co-variations are within the normal variability observed in relatively pristine mangrove sites, and most of the observed differences between Mtoni and Ras Dege are likely to be due to their contrasting environmental settings (e.g. salinity regime, tidal amplitude, forest/open water ratio, …). Data collected during diurnal cycles indicate the dominance of tidal effects on the biogeochemical signatures in the water column in the pristine site, and a tight coupling with processes occuring in the intertidal areas, e.g. through porewater drainage. For organic and inorganic carbon, this would imply that the variability in the sources and concentrations of these C pools is large enough to mask any additional inputs derived from other sources such as sewage inputs – such inputs may exist but biogeochemical signatures of these carbon pools per se are insufficient to trace such inputs. Distinct differences in inorganic N profiles (NH4+ and NO3-) were found and consistent with expectations, but data from other sites suggest that nutrient concentrations in itself are not necessarily a diagnostic for anthropogenic inputs. In the impacted site, photosynthesis by dense mats of benthic microalgae is sufficiently high to be traced in the water column oxygen levels and δ18O signatures, and stable isotope data on bacterial biomarkers suggest an exceptionally high dependency of sediment microbial communities on benthic primary production -at least in the surface layers. Finally, distinctly higher δ15N signatures were found in suspended matter in the impacted site, which is likely to be closely linked to anthropogenic nutrient inputs – δ15N signatures of different biological compartments therefore appear to be a potentially powerful and integrative measure of nutrient loading at the ecosystem level.
