The estimate of the surface conductance for water vapor transfer typical of a primary forest has been carried out, near Manaus, using two different methods based on regression equations which try to associate this conductance to distinct physical variables. The equation proposed by Allen (1986) was established through the correlation with the daily mean net radiation. The estimate proposed by Roberts et al. (1990) tried to correlate the stomatal conductance to the specific humidity deficit and to the solar radiation, using hourly mean data, and in this instance these values have been used to calculate the corresponding daily mean. A comparison of the estimated results using these methods showed the first one (Allen) with estimates always lower than the second one (Roberts), of about 30%(±8) on the average. The deviations of both methods have also been relatively different, and although Roberts has shown higher conductance indices, the deviations were of lesser magnitude than those of Allen, being these values more stable. The lowest differences have been observed during the dry period, mainly in the months of August, September, and October, when they were within the limit determined by the sum of the standard deviations of both of them.
evapotranspiration; conductance; primary forest; Amazonia