Micrometeorological heat, water, and CO2fluxes have been widely monitored with the eddy covariance (EC) method. Its uncertainty information should be as important as the flux measurement itself and is indispensable to studies addressing site intercomparison, model validation, and model-data synthesis. Therefore, we estimated the fractional uncertainty ϕ using EC data collected over various types of land cover and at different instrumental heights. We also investigated ϕ characteristics according to a spatiotemporal scale and the flux averaging interval τ. As a result, we suggest that if ϕ is estimated under the EC measurement conditions satisfying stationarity, it is stable and uniform regardless of the land cover, spatiotemporal scale, and kind of flux. Based on the constancy of ϕ, we determined the baseline as a function of τ for a stationarity or a heterogeneity index of the EC measurement. © 2011, The Society of Agricultural Meteorology of Japan. All rights reserved.