“…The work presented here is based on empirical relationships from global datasets, but may not be applicable to certain environments. For example, studies have shown that within fog-dominated ecosystems (Anchukaitis et al, 2008;Roden et al, 2009), semiarid environments (Kanner et al, 2014), or regions where seasonal variation in d 18 O MW is affected by large-scale vapor transport patterns (Bowen, 2008;Bowen and Wilkinson, 2002;Breecker, 2013;Liu et al, 2013;Saurer et al, 2012) the local effects of changes in moisture source on the intra-annual d 18 O cell pattern may be important. Our dataset does include such ecosystems [e.g., site 26 (significant fog); sites 1, 4, 11e13 (dry, MAP < 500 mm); sites 6, 16, 18e19, 22e25, 30e31 (monsoonal)] and our model using only changes in seasonal temperature and precipitation amount does successfully explain much of the variability in the measured intraring d 18 O cell signal worldwide (R 2 ¼ 0.82) (Fig.…”