Factors controlling Arctic denitrification

Denitrification of the Arctic winter stratosphere has been calculated using a 3-D micro-tion is assumed to occur through the sedimentation of low number concentrations of large nitric acid trihydrate (NAT) particles, as observed extensively in 1999/2000. We 5 examine whether the meteorological conditions that allowed NAT particles to grow to the very large sizes observed in 1999/2000 also occurred in the other cold winters. The results show that winter 1999/2000 had conditions that were optimum for deni-trification by large NAT particles, which are a deep concentric cold pool and vortex. Under these conditions, NAT particles can circulate in the cold pool for several days, 10 reaching several micrometres in radius and leading to a high downward flux of nitric acid. The other winters had shorter periods with optimum conditions for denitrification. However, we find that NAT particles could have grown to large sizes in all of these winters and could have caused significant denitrification. We define the quantity " closed flow area " (the fraction of the cold pool in which air parcel trajectories can form closed 15 loops) and show that it is a very useful indicator of possible denitrification. We find that even with a constant NAT nucleation rate throughout the cold pool, the average NAT number concentration and size can vary by up to a factor of 10 in response to this meteorological quantity. These changes in particle properties account for a high degree of variability in denitrification between the different winters. This large mete-20 orologically induced variability in denitrification rate needs to be compared with that which could arise from a variable nucleation rate of NAT particles, which remains an uncertain quantity in models.