Staggered-set clearcutting on Pacific Northwest Federal lands has created a fragmented landscape of different age Douglas-fir forests ranging from early seral (0–15 years old), young (15–80 years), intermediate (80–200 years), mature (200–400 years), to old-growth (>400 years).  While early seral stands can compromise up to 40% of the total forest coverage in the Western Cascade Mountains and are an essential component of the regional carbon budget, ecosystem exchange within this youngest age class has not been thoroughly studied with eddy covariance.  Early seral stands are usually fetch-limited which creates a unique set of micrometeorological concerns for the eddy covariance technique.  Here we present CO₂, H₂0 and energy flux and meteorological data from two early seral stands and the Wind River AmeriFlux old-growth forest during two growing seasons (March–October).  We show an alternative approach to the usual friction velocity (u_*) method for determining periods of adequate atmospheric boundary layer (ABL) mixing based on the ratio of mean horizontal or vertical wind speed to a modified turbulence kinetic energy scale (u_TKE).  This new parameter was used to identify half-hour fluxes measured during conditions when transport by mean flow could no longer be neglected compared to turbulent flow in the wind field.  We observed fundamental differences between u_* and u_TKE screening depending on the time of day and site, with u_TKE being the more conservative parameter (i.e., u_TKE screening eliminated a higher percentage of flux data).  The screening parameter, u_TKE, in addition to footprint modeling, showed that 80% of nighttime fluxes at the early seral stands were measured during inadequate turbulence and fetch conditions.  Daytime CO₂ fluxes in small clear-cuts, on the other hand, were accurately measured as long as the micrometeorological conditions were carefully evaluated.  Midday CO₂ fluxes peaked annually in April at the old-growth stand and were -14.0 ± 3.4 μmol m² s^-1 in 2006 and -12.3 ± 2.1 μmol m² s^-1 in 2007.  In contrast, midday CO₂ fluxes were significantly (P < 0.0001) smaller (less net carbon uptake) at the early seral stands in April and averaged -4.0 ± 1.4 μmol m² s^-1.  Peak midday CO₂ fluxes were observed two to three months later at the younger stands: maximum CO₂ fluxes were between -10.2 ± 2.0 μmol m² s^-1 and -8.7 ± 0.9 μmol m² s^-1 in June while net CO₂ uptake decreased sharply at the old-growth forest and continued to decline throughout the summer in both 2006 and 2007.  We conclude that a velocity scale based on turbulence kinetic energy, instead of friction velocity, is a more accurate way of assessing ABL and stable boundary layer mixing and therefore, net ecosystem exchange, because it directly indicates turbulence generated by buoyancy as well as from mechanical forces, instead of being tied only to the momentum flux.