Spatial heterogeneity and lake morphology affect diffusive greenhouse gas emission estimates of lakes

Most estimates of diffusive flux (F) of methane (CH₄) and carbon dioxide (CO₂) from lakes are based on single-point flux chamber measurements or on piston velocity (k) modeled from wind speed and single-point measurements of surface water gas concentrations (C_aq). We analyzed spatial variability of F of CH₄ and CO₂, as well as C _aq and k in 22 European lakes during late summer. F and k were higher in the lake centers, leading to considerable bias when extrapolating single-point chamber measurements to whole-lake estimates. The ratio of our empirical k estimates to wind speed-modeled k was related to lake size and shape, suggesting a lake morphology effect on the relationship between wind speed and k. This indicates that the error inherent to established wind speed models can be reduced by determining k and C_aq at multiple sites on lakes to calibrate wind speed-modeled k to the local system. Key Points Diffusive greenhouse gas flux from lakes has a distinct within-lake distribution Single-point estimates can lead to biased diffusive gas flux estimates The relationship of wind speed at 10 m with piston velocity varies within lakes