Title: Quantification of nitrous oxide emissions from rivers and streams in Mara basin, Kenya

Author: Clarisse Ishimwe

Supervising Institution: IHE Delft  - Institute for Water Education

Year: 2020



N2O is a powerful greenhouse gas (GHG) with a global warming potential 298 times that of carbon dioxide (CO2). Stream and rivers have been reported to contribute a significant amount of N2O emissions globally. However, data on N2O emissions from tropical rivers in Sub – Saharan Africa is still scarce, despite the disproportionately large contribution from tropical rivers to global N2O emissions. To partly address this, my study aimed to quantify N2O concentrations and fluxes from a tropical African river, while accessing several of their controlling factors such as stream discharge, water quality variables, stream order and catchment land use. Additionally, I also accessed how daytime measurements alone could bias the daily N2O fluxes, through diel sampling of the emissions. In total, 58 sites were sampled in the upper and middle sub-catchments of the Mara River for N2O concentrations, water quality parameters, and hydrological parameters. The streams and rivers drained three main land use categories i.e cropland, forest and livestock and were of stream orders from 1 to 8. Overall, N2O concentrations ranged from 0.005 to 0.14 μmol L-1 with fluxes ranging from -53.92 to 1171.203 μmol m-2d-1. Stream order and discharge did not significantly account for the variation measured in the N2O concentration and fluxes. However, fluxes and concentrations were significantly controlled by land use. Cropland streams had significantly higher mean fluxes of 240.26 ± 31.54 μmol m-2d-1 compared to livestock streams with a mean value of 64.46 ± 8.05 μmol m-2d-1. However, fluxes did not differ significantly between cropland and forested streams that had a mean value of 152 ± 31.54 μmol m-2d-1 and 137.12 ± 34.25 μmol m-2d-1, respectively. Based on the differences in N2O concentrations amongst land uses, I inferred to different controlling processes for each land use. N2O concentrations in cropland stream was positively correlated to NO3 while forested streams was negatively correlated with dissolved oxygen and pH. For livestock streams no correlation with any physical chemical parameters. Based on these relationships I hypothesized that the main biogeochemical process in our system was nitrification. The diel variation results showed that there was no difference in N2O fluxes in day- time. We concluded that N2O was varying across different land use, and the stream order did not define the change observed in N2O concentration and fluxes.