Spatial Variability of Greenhouse Gas Emissions from Tropical Reservoirs

Abstract: Hydroelectricity has for a long time been considered climate neutral due to it being a renewable source of energy. During the last years however, studies have shown that emission magnitudes from hydroelectric reservoirs may be equal to those of fossil fuel power plants. Reservoir emissions are largest in tropical regions, where CO2 diffusion and CH4 ebullition are main contributors to the overall emission rate. It is also in tropical regions where extraction of hydroelectricity is expected to experience a sharp rise in coming years. In a study published 2011 it was hypothesised that previous estimates have completely missed ebullition hotspots in reservoirs, and thereby underestimated CH4 emission by at least one order of magnitude.   Spatial variability of CH4 ebullition rates has been estimated for two tropical reservoirs: nutrient-poor Chapéau d’Uvas Reservoir (CDU) and nutrient-rich Funil Reservoir (FUN). Spatial variability of diffusion and the total emission rate has furthermore been estimated for CDU. Additionally, two methods used for measuring gas transfer rates (a parameter important for explaining diffusion) have been compared.   The obtained estimate in total emission rate was 3,094 mg CO2-eq m-2 day-1, which amounts to half of the average in the most recent global assessment for tropical reservoirs. The estimation of ebullition emission in FUN was 4,000 times lower than in CDU, likely due to a higher rate of increase in hydrostatic pressure during sampling in FUN. Similarities identified between CH4 ebullition, CO2 diffusion and CO2 concentration were: generally higher rates in bays and in the main inflow than in the main reservoir area. No statistically significant differences in spatial variability of ebullition between inflow areas and non-inflow areas were identified. The method-comparison for gas transfer rate measurements indicated that the discrepancy between the methods increased with higher average values measured.