The eddy covariance technique is the most common method of measuring forest evapotranspiration directly. Evapotranspiration can also be estimated by solving for the residual of the energy balance measurements of net radiation, sensible heat flux density and ground heat flux density. However, the lack of energy balance closure of all components creates uncertainty in using the residual for operational estimates of evapotranspiration. Here, we show that the residual can agree well with the direct eddy covariance measurement after data have been quality-controlled and gap-filled. The technique is tested using energy balance data from three young, post-fire forests in Saskatchewan Canada during both wet and dry years. The seasonal cumulative evapotranspiration agrees within less than 5% and weekly agreement is very good. There is more scatter among the daily data and we do not recommend this method for shorter than daily averages unless all energy storage terms are measured very well. We conclude that forest evapotranspiration can be estimated cost effectively using the energy balance residual technique as daily, weekly and seasonal averages.