We compare in this paper direct measurements obtained on the towers on San Salvatore Mountain (Switzerland) and on the Gaisberg Mountain (Austria). They are situated in similar topographical environments but in different lightning activity zones. Direct measurements of lightning currents on these towers have revealed a major difference in terms of the number of downward flashes. While measurements made by Berger and co-workers revealed a significant number of downward flashes on the two towers on San Salvatore Mountain, more recent observations on the Gaisberg and Peissenberg towers were essentially composed of upward flashes. We use in this paper a new method to estimate the proportion of upward/downward flashes to a given tower, based on the data from lightning location systems. The analysis using the proposed method explains the discrepancy in terms of the measured number of downward flashes in the Gaisberg and San Salvatore towers. The analysis presented reveals also that in the evaluation of the percentage of upward flashes initiated from a tall structure, different parameters should be carefully examined, namely (i) the value of the ground flash density, (ii) the topographical conditions, and (iii) the presence of other tall structures in the region from which upward flashes might be initiated. (c) 2012 Elsevier B.V. All rights reserved.