Reply to comment by Röckmann and Kaiser on ‘‘Evidence for O-atom

[1] Based upon the authors’ questioning of the existence of the C2v intermediate, we have reviewed our evidence for the existence of this state. It now appears that this state was in fact an artifact of our calculation [Yung et al., 2004], and was a saddle point rather than a true minimum. Our desire to provide a timely response to this criticism has kept us from determining exactly what minimum structure will be obtained by a full minimization at the level of theory employed. However, it is clear that the C2v symmetry of the compound is broken in such a way that the two N-O bonds are no longer equivalent. We are grateful to the authors for helping us resolve this issue. [2] We further note that new laboratory studies of the O(D) + N2O reaction [Nishida et al., 2004] show that the inelastic channel (O(D) + N2O ! O(P) + N2O) is about 5% that of the reactive channels (2NO and N2+O2). Based on the facts that CO2 and N2O are isoelectronic linear triatomic molecules and that their overall rates for the reaction with O(D) are comparable, it is reasonable to assume that the nominally ‘non-reactive’ reactions in the two systems have similar probabilities for oxygen atom isotopic exchange. By analogy with O(D) + CO2 reactions [Perri et al., 2003], the branching ratio of the elastic channel for (O(D) + N2O ! O(D) + N2O) is suggested to be 2.5%, so that the ratio of the exchange to reactive channels, g, is <10%. With g = 10%, DO is limited to a value of about 0.1% at dO = 20% because its magnitude is determined by both the chemical exchange and transport times. With these reaction rate coefficients, the O(D) + N2O reaction cannot be the dominant source of the oxygen anomaly in stratospheric N2O, but further laboratory experiments are clearly needed to determine the total branching ratio of the exchange channels. Additional upper bounds on the magnitude of g may well be obtained from a careful analysis of the experiments of Johnston et al. [1995] and Kaiser et al. [2002], but we have not yet completed this analysis. It is important to note that we have omitted the doubly substituted reactions since their probability is negligible compared with the effects addressed in the paper. An analysis of g on the existing data is shown by Röckmann and Kaiser [2005].