Nitric acid in the stratosphere based on Odin observations from 2001 to 2007 ? Part 2: High-altitude polar enhancements

The wintertime abundance of nitric acid (HNO 3 ) in the polar upper stratosphere dis-plays a strong inter-annual variability, and is known to be strongly inﬂuenced by energetic particle precipitation, primarily during solar proton events, but also by precipitating electrons in the auroral zone. While wintertime HNO 3 enhancements in the 5 polar upper stratosphere had been occasionally observed before, from the ground or from satellite, we present here measurements by the Sub-Millimeter Radiometer instrument aboard the Odin satellite through 6 full annual cycles (2001 to 2007). Major solar proton events, e.g. during November 2001 or the Halloween solar storms of autumn 2003, lead to a two-stage HNO 3 enhancement, likely involving di ﬀ erent chemical 10 reactions: a fast (about 1 week) in-situ enhancement from the mid to the upper stratosphere is followed by a slower, longer-lasting one, whereby anomalies originating in the upper stratosphere can descend within the polar vortex into the lower stratosphere. We highlight the fact that the actual chemical coupling between the upper and lower atmosphere involves a complex interplay of chemistry, dynamics and energetic particle 15 precipitation.

polar upper stratosphere had been occasionally observed before, from the ground or from satellite, we present here measurements by the Sub-Millimeter Radiometer instrument aboard the Odin satellite through 6 full annual cycles (2001 to 2007). Major solar proton events, e.g. during November 2001 or the Halloween solar storms of autumn 2003, lead to a two-stage HNO 3 enhancement, likely involving different chemical reactions: a fast (about 1 week) in-situ enhancement from the mid to the upper stratosphere is followed by a slower, longer-lasting one, whereby anomalies originating in the upper stratosphere can descend within the polar vortex into the lower stratosphere. We highlight the fact that the actual chemical coupling between the upper and lower atmosphere involves a complex interplay of chemistry, dynamics and energetic particle 15 precipitation.

Introduction
Nitric acid (HNO 3 ) is a key minor constituent of the middle atmosphere, part of the odd nitrogen family (NO y ), and a reservoir for the active nitrogen species (NO x ), which provide a major ozone loss catalytic cycle in the middle and upper stratosphere. In the 20 lower stratosphere, HNO 3 plays a multi-facetted role in ozone depletion.
Stratospheric HNO 3 is produced through gas phase reaction of hydroxyl (OH) with NO 2 , and heterogeneous chemical conversion of N 2 O 5 , the latter constituent being produced by gas phase reactions in the cold polar night conditions and easily thermally decomposed. The HNO 3 sinks are photo-dissociation and reaction with OH. Its photo-stratosphere, HNO 3 has hence a pronounced seasonal cycle in the polar regions. The HNO 3 -rich layer peaks at around 25-30 km at high latitudes in winter. Another sink of HNO 3 is the sequestration from the gas phase during polar stratospheric cloud formation, and its irreversible removal through sedimentation (denitrification).
Stratospheric HNO 3 has been observed by means of ground-based, balloon and 5 aircraft and satellite instrumentation. The most complete dataset to date has been provided by the Microwave Limb Sounder (MLS) instrument aboard UARS (Santee et al., 2004), albeit not in the upper stratosphere. While the first part of this article (Urban et al., 2008) shows 6 annual cycles (2001)(2002)(2003)(2004)(2005)(2006)(2007) of satellite observations of stratospheric HNO 3 by the "Odin Sub-Millimetre Radiometer" (SMR), we focus in this second part 10 on describing the high-altitude polar enhancements, which have not been documented over so many years by other satellite instruments. In winter, enhanced layers of HNO 3 are commonly observed at high altitudes in the polar regions, as revealed by groundbased (de Zafra and Smyshlaev, 2001) or satellite observations (Austin et al., 1986;Lopez-Puertas et al., 2005b, hereafter LP05;Orsolini et al., 2005, hereafter OR05). 15 These enhanced layers appear first in the uppermost stratosphere and tend to descend within the winter polar vortex. While recurrent, these enhancements vary widely in amplitude from year to year. Exceptionally strong enhancements have been linked to energetic particle precipitation (EPP) events and anomalous descent of mesospheric air.

20
Stratospheric NO x abundances are amplified in-situ during the strongest solar proton events (SPEs) (also known as the NO x direct effect), or through downward transport of mesospheric air (also known as the NO x indirect effect), enriched in NO x by EPP, i.e. SPEs or low energy electron precipitation from auroral activity. Various EPP events have led to upper-stratospheric NO 2 abundances over a hundred ppb ( spheric Sounding (MIPAS) observations in particular, revealed a two-stage evolution of HNO 3 enhancements (LP05; OR05). Immediately following the SPE, a short-lived (about 1 week) stratospheric layer of enhanced HNO 3 , peaking at 2-2.5 ppb, was observed above 35 km by LP05, who suggested formation through gas phase reactions with the enhanced OH abundance, and through ion chemistry in darkness. Newly 5 reprocessed MIPAS retrievals indicate that the HNO 3 enhancements extend into the upper stratosphere and lower mesosphere (Lopez-Puertas, 2007). A weaker shortlived increase in HNO 3 was also observed in the southern hemisphere. Incidently, these short-lived chemical perturbations were not limited to NO 2 and HNO 3 . MIPAS observed perturbed chlorine family species (ClO, HOCl) (von Clarmann et al., 2005), 10 and O 3 depletion was observed by a variety of satellite instruments and confirmed by model studies (Lopez-Puertas et al., 2005a;Seppalla et al., 2004;Rohen et al., 2005;Jackman et al., 2005Jackman et al., , 2007. In a second stage, several weeks after the SPE, an anomalous HNO 3 -rich layer was first observed at about 45 km (OR05), and intensified considerably while descend- 15 ing confined in vortex air. By mid-January, it has reached 30 km and vortex-averaged HNO 3 abundances were as high as 13-15 ppb, leading to double-peaked high-latitude HNO 3 profiles. Several mechanisms have been proposed to explain these long-lasting high-altitude HNO 3 enhancements, which invoke heterogeneous reactions converting N 2 O 5 into HNO 3 . de Zafra and Smyshlaev (2001) followed earlier suggestions 20 (Bohringer et al., 1993) that hydrated ion clusters might be the seat of such heterogeneous reactions, albeit sulphate aerosols were suggested to play a role below 35 km (Bekki et al., 1995). This second stage requires a large downward flux of NO 2 to generate N 2 O 5 , but also a high degree of vortex confinement. The background abundance of hydrated ion cluster is thought to be generated by galactic cosmic rays.

25
One might call the first stage of HNO 3 enhancements, the fast effect, and the second stage the delayed effect. While the HNO 3 enhancements share some characteristics of the NO x enhancements, such as high-altitude origin, polar confinement and descent, they do not need to follow or coincide with them. Late winter or spring NO x pulses, Introduction  (Randall et al., 2005(Randall et al., , 2006, might not give rise to HNO 3 enhancements, as the slow ion cluster chemistry requires a stable vortex, and coldness and darkness to build up HNO 3 , conditions which are not provided close to the stratospheric final warming. These NO x pulses nevertheless were important for the upper stratospheric ozone budget, leading for example to nearly 5 60% ozone destruction at 45 km in spring 2004 (Natarajan et al., 2004;Randall et al., 2005). The aim of this paper is to show high-altitude HNO 3 polar enhancements, including the occasional two-stage time-development, in 6 annual cycles of Odin/SMR satellite observations from 2001 to 2007. The companion paper describes the characteristics Winter 2003/2004. The two-stage development first seen in MIPAS (LP05; Lopez-Puertas, 2007; OR05) is confirmed by the SMR data. During the first stage (fast effect), which extends from above 1300 K into the upper stratosphere-lower mesosphere, the 20 mixing ratios anomalies are smaller than the 2-2.5 ppb in MIPAS data (near 45 km), but are reduced due to time smoothing applied here. In the second stage, the mixing ratios of about 12 ppb at 960 K in early January 2004 are in good agreement with MIPAS.
Winter 2004/2005. NH anomalies during the winter 2004/2005 appear more complex to interpret, as two "streaks" of enhancements are observed, one starting early in  cember which could be interpreted as due to the normal early winter descent, while the second, deep enhancement coincides with the strong SPE of mid-January, and hence could be interpreted as a fast effect. The latter SPE ranked as number 11 of the last 9596 Introduction from the mesosphere into the polar upper stratosphere provides NO x for HNO 3 heterogeneous conversion, this amount is highly variable, and strongly influenced by EPP. Exceptional enhancements require a large source of NO x in the form of a strong SPE or an anomalously high auroral activity. Next, dynamics plays a strong role in channelling the descending chemical anomalies. An enhanced descent well-confined in the 5 strong vortex would be leading to large, NO x stratospheric enhancements, which are a prerequisite for the HNO 3 build-up. Stratospheric sudden warmings could also act to damp the HNO 3 anomalies by bringing air in sunlit regions, and alter the vertical mixing by gravity waves. Not only the magnitude but also the seasonal timing of EPP events is important. with the main layer when the abundance is still high, as clearly seen during the two strongest episodes in 2002 and 2004. In the SH, on the contrary, the descending layer normally reaches the lower stratosphere when the main layer abundance has already decreased considerably. A further point to note is that, as the high-altitude layer descends, the mixing ratios increase indicating continued production.

20
The ODIN/SMR HNO 3 observations provide for the first time a multi-year record of polar enhancements at high altitudes, and their downward propagation inside the winter polar vortex. Outstanding enhancements are seen during previously studied EPP or strong mesospheric descent events (such as the Halloween storms in late autumn 2003, or the austral winter 2003), but also during more recent ones. The two-stage effect), characterised by much higher anomalies, up to 10-15 ppb range in unfiltered data. The ODIN/SMR HNO 3 observations nicely show the merging of the descending layer originating at high altitudes with the main layer near 25 km after the seasonal peak.
The descending low or high anomalies appear somewhat analogous to the tropical 5 "tape-recorder" effect, that describes how low-latitude tracer anomalies imprinted at the tropopause level ascend over years, keeping a memory of their initial composition, and giving rise to layered anomalies in the tropical stratosphere. In this case, it is acting at high latitudes, and in reverse (propagating downwards) fashion, and on a faster (seasonal) scale: HNO 3 anomalies are imprinted near the stratopause, descending 10 to the lower stratosphere over the course of the winter, giving rise to layering. The amplitude of the anomalies also increases with time, i.e. during descent, unlike the tape-recorder effect.
Further modelling studies are needed as a step toward implementing appropriate schemes to represent these processes affecting the stratospheric NO y budget into 15 global chemical transport models.