Inland waters, particularly small and shallow lakes, are significant sources of carbon dioxide (CO2) and methane (CH4) to the atmosphere. However, the spatial in-lake heterogeneity of CO2 and CH4 production processes and their drivers in the sediment remain poorly studied. We measured potential CO2 and CH4 production in slurry incubations from 12 sites within the small and shallow crater lake Windsborn in Germany, as well as fluxes at the water–atmosphere interface of intact sediment core incubations from four sites. Production rates were highly variable and ranged from 7.2 to 38.5 µmol CO2 gC−1 d−1 and from 5.4 to 33.5 µmol CH4 gC−1 d−1. Fluxes ranged from 4.5 to 26.9 mmol CO2 m−2 d−1 and from 0 to 9.8 mmol CH4 m−2 d−1. Both CO2 and CH4 production rates and the CH4 fluxes exhibited a significant and negative correlation (p<0.05, ρ<−0.6) with a prevalence of recalcitrant organic matter (OM) compounds in the sediment as identified by Fourier-transformed infrared spectroscopy. The carbon / nitrogen ratio exhibited a significant negative correlation (p<0.01, ρ=−0.88) with CH4 fluxes but not with production rates or CO2 fluxes. The availability of inorganic (nitrate, sulfate, ferric iron) and organic (humic acids) electron acceptors failed to explain differences in CH4 production rates, assuming a competitive suppression, but observed non-methanogenic CO2 production could be explained up to 91 % by prevalent electron acceptors. We did not find clear relationships between OM quality, the thermodynamics of methanogenic pathways (acetoclastic vs. hydrogenotrophic) and electron-accepting capacity of the OM. Differences in CH4 fluxes were interestingly to a large part explained by grain size distribution (p<0.05, ρ=±0.65). Surprisingly though, sediment gas storage, potential production rates and water–atmosphere fluxes were decoupled from each other and did not show any correlations. Our results show that within a small lake, sediment CO2 and CH4 production shows significant spatial variability which is mainly driven by spatial differences in the degradability of the sediment OM. We highlight that studies on production rates and sediment quality need to be interpreted with care, though, in terms of deducing emission rates and patterns as approaches based on production rates only neglect physical sediment properties and production and oxidation processes in the water column as major controls on actual emissions.