LAUSR

Abstract The failure of the Monin–Obukhov Similarity Theory (MOST) in the roughness sublayer is a major problem for the estimation of fluxes over tall forests, whenever indirect methods that rely on MOST, such as flux-gradient or the variance method, are involved. While much research focuses on micrometeorological measurements over temperate-climate forests, very few studies deal with such measurements over tropical forests. In this paper, we show evidence that some similarity functions over the Amazon forest are somewhat different from temperate forests. Comparison of the nondimensional scalar gradients canonical values for the inertial sublayer with our measurements in the roughness sublayer showed smaller deviations than what is usually reported for temperate forests. Although the fluxes of water vapor and CO2 derived from mean profiles show considerable scatter when compared with the eddy covariance measurements, using calibrated dimensionless gradients it is possible to estimate their mean daily cycle during the period of measurement (36 days in May and June, transition between rainy and dry season). Moreover, since mean ozone profiles were available, although without the corresponding eddy covariance measurements, mean daily ozone fluxes were calculated with the flux-gradient method, yielding a nighttime value of −0.05 and a daily peak of −0.45 μg m−2 s−1 (−1.04 and −9.37 nmol m−2 s−1, respectively). These values are comparable to previously measured fluxes in the literature for the Amazon forest.