Comment on "Reassessment of rain experiments and operations in Israel including synoptic considerations" by Zev Levin, Noam Halfon and Pinhas Alpert [Atmos. Res. 97 (2010) 513-525]

Levin et al. (referred here as LHA) made a sweeping conclusion that cloud seeding has been ineffective in Israel. They claimed that the results of the Israel-2 experiment could be fully ascribed to synoptic bias. However, the cross over analysis of Israel-2 has shown the same already in 1990 by Gabriel and Rosenfeld. While LHA only showed differences in the 850. hPa winds being stronger in the north seeded days of Israeli-2, Rosenfeld and Farbstein had already shown quantitatively in 1992 that this synoptic bias explains less than half of the indicated seeding effect in the north. Furthermore, accepting that a cross over design protects against synoptic bias means that the highly positive and statistically significant indicated seeding effect in Israeli-1, which was based on a cross-over design, cannot be explained by such synoptic bias. Instead, LHA completely ignored the Israeli-1 experiment.LHA's evaluation of the operational seeding used implied assumption that amounted to historical comparisons that have already been shown to be invalid, especially for the Israeli situation due to the decreasing trend of target-control ratio, especially over the eastern upper Galilee. For all of the above reasons, the conclusion of LHA that cloud seeding is ineffective in Israel is not supported by the data.Given the uncertainties, both physical and statistical, the Israeli water authority has embarked on the Israeli-4 randomized cloud seeding experiment, which is aimed at testing the hypothesis that cloud seeding might be affecting mainly the precipitation in the orographic clouds developing over the hills in the catchment of the Jordan River.