A procedure, recently proposed by the authors to calculate actual evapotranspiration on the basis of parameterization of the atmospheric boundary layer, was applied for the period April through November 1969–1972 with upper air (‘rawinsonde’) meteorological data from Omaha, Nebraska. The accuracy of the 1‐day, 3‐day, weekly, and monthly evapotranspiration totals was assessed by comparing them to evapotranspiration data estimated by a suitable adjustment of potential evapotranspiration from a nearby catchment in the Treynor basins. Several alternative methods for determining the height of the boundary layer and the wind speed at the top of the boundary layer were used; as a result, seven variations of the basic procedure were tested. The similarity functions for water vapor D were found to be smaller than the corresponding functions for sensible heat C and not equal as was previously assumed. Difficulties were encountered, since published upper air humidity data were in error; nevertheless, the method was reasonably successful on a monthly basis with a correlation coefficient of the order of 0.83 but far less so on a daily basis. The recommended version of the method uses the height of the temperature inversion as an estimate of the boundary layer height and the measured wind as the estimate of the wind speed at the top of the boundary layer.