Accuracy levels of measurements with the Piche evaporimeter as an ancillary anemometer for the interpolation and/or extrapolation of wind speed levels, in combination with electrical cup anemometers, were determined. Data from the shaded coffee in Lyamungu (Tanzania) showed that this procedure had more than 10% accuracy for daily average levels at individual experimental sites, as long as wind speed did not drop too often below 0.75 m s-1. For short runs, this accuracy was noticeably lower but improved to the same order of magnitude when using wind speed reduction ratios of the two types of anemometers. Such improvements were confirmed by the data taken in a Savanna woodland in Setchet (Tanzania) and in a complex agroforest system with hedges and intercropped trees in Matanya (Kenya). However, within a multiple shelterbelt system at Yambawa (Nigeria), under high advective conditions and high gradients of advection between the belts, the Piche system was generally not usable, and neither were ancillary anemometers with high variations during rainy season. Only homogeneity of permeability of the belts could be well determined with the Piches. Earlier results obtained in a shelterbelt in Sehaimab (Sudan) proved that application of a reference for the agroforest systems was more accurate than using a reference outside them, which was confirmed by the data from Kenya and to a certain extent also by the data obtained in Nigeria.