The objective of this work was to estimate Angström-Prescott equation mean coefficients in Brasília, DF, Brazil, and their effects on the calculations of global radiation (Rs) and reference crop evapotranspiration (ETo). It was used meteorological data from Embrapa Hortaliças weather station, from 1978 to 2003. The mean coefficients were determined by linear regression analysis considering monthly, annual and total periods. Using the calculated coefficients for different periods and those provided by FAO (0.25 and 0.50), Rs and ETo were estimated and their performances were evaluated. Monthly mean coefficient a varied from 0.241 to 0.345, and b from 0.430 to 0.515. Annual mean coefficients a and b were 0.278 and 0.498, and total means were 0.282 and 0.490, respectively. These coefficients did not differ from each other on radiation estimates. The month of July provided the worse estimates and performance. When using calculated Rs estimates with monthly, annual, total and FAO coefficients, estimated values of ETo did not differ from observed ones, providing excellent performance, despite the mean coefficient used. Depending on the final application, it is suggested the use of total mean coefficients due to their easiness to estimate the shortwave radiation.
energy balance; actual duration of sunshine; FAO-Penman-Monteith
