Methane and carbon dioxide, which are two important greenhouse gases, are products of organic matter decomposition in the aquatic system. Both gases are transferred to the atmosphere mainly through ebullition and diffusion. Diffusion often accounts for the majority of CO₂ transport and ebullition for the CH₄ transport, as methane is a less soluble gas and is formed mostly under anoxic conditions, as found in bottom sediments of reservoirs. Ebullition process is highly variable and depends on physical, chemical, and biological conditions. As a result, the quantification and prediction of GHG emissions from water bodies are challenging, adding to the fact that the relationship of ebullition with the interplay of environmental conditions is still not well established. This research aims to improve the process-based understanding of the spatial and temporal variability of ebullition in aquatic systems. The study site is Passaúna reservoir, which is a water supply reservoir managed by Sanepar and is being monitored within MuDak-WRM project (German-Brazilian project). The ebullition fluxes are continuously recorded at three locations of the reservoir by automated bubble traps while additional parameters as dissolved oxygen, water temperature, and current velocity were measured in a fixed platform at one location. Meteorological information is available from two nearby weather stations. So far, more than 645 days of ebullition flux was recorded with the highest emissions occurring during warm months. The correlation between ebullition flux and isolated variables results in very low correlation coefficients (Spearman’s coefficients range: -0.07 ≤ r_s ≤ 0.44), as the process depends on the interplay of multiple variables. It was used a random forest algorithm to assess the relative importance of 6 variables to the ebullition flux. For the hourly time scale the barometric pressure was the most relevant parameter, followed by water level, whereas the least important was dissolved oxygen. Spatially the deepest location in the reservoir has the lowest emissions, nevertheless, additional analysis will be conducted to better describe the spatial variability of fluxes. The following steps of this research will be to combine spatio-temporal data like hydro acoustic sediment information to have an integrated analysis of ebullition from the reservoir.