Cosmic-ray–induced ionization can lead to essential chemical changes in the polar stratosphere with enhanced production of “odd nitrogens” NOy (e.g., N, NO, NO2, NO3, 2N2O5, BrONO2, ClONO2, HO2NO2, and HNO3) (see, e.g., Jackman et al., 1990, 1993; Vitt et al., 2000). Due to ionization of the ambient air and subsequent dissociation of O2 and N2, energetic particles precipitating into the atmosphere facilitate formation of the odd nitrogen. The abundance of odd nitrogen dramatically responds to variations in the SEP flux on the background of a smooth 11-yr cycle due to GCR variations (Thomas et al., 2007). Odd nitrogen is long-lived during the polar night, and some species, such as nitric acid HNO3, can be effectively transported down and finally stored in polar snow/ice. By measuring such chemical species in a stratified polar-ice archive, one can obtain a record of the physical-chemical conditions of the polar atmosphere (Zeller et al., 1986; Zeller and Dreschhoff, 1995). Under some basic reasonable assumptions (stable atmospheric properties, no mixing in snow, etc.), the abundance of compounds, related to the stratospheric chemistry such as nitrate ions (NO), provides a unique and long-term (potentially up to 105 years) record of the radiation environment and climate in polar atmosphere. The concentration of nitrates has been measured in polar ice from both the Southern (South Pole, e.g., Dreschhoff and Zeller, 1990) and Northern (Greenland, e.g., Zeller and Dreschhoff, 1995; Dreschhoff and Zeller, 1998) polar caps. The Greenland series of ultra-high resolution (better than 30 equispaced samples per year) provides a unique record in which (groups of) individual strong SEP events can be identified for as far back as 1562.
The nitrate abundance measured in polar ice depicts a clear seasonal cycle with a peak during the local summer because of increased snow sublimation under sunlight (Zeller et al., 1986). Moreover, the nitrate series contains clear time stamps of major volcano eruptions, which are apparent in the measured signal. Altogether it provides a solid basis for absolute dating of the samples with a time resolution within one year (McCracken et al., 2001b). Thus, the nitrate concentration in well-dated polar-ice cores provides a unique opportunity to evaluate the flux of SEP in the past, before instrumental observations. High time resolution allows the separating of (groups of) individual SEP events, such as the famous Carrington event in September 1859 (Shea et al., 2006; Thomas et al., 2007). It has been suggested that nitrate enhancements caused by strong SEP events can be reliably distinguished both from the relatively slow and shallow variations of galactic cosmic rays and from meteorologically-derived nitrate peaks.
The first studies based on the correlation between peaks of nitrate concentration and measures of solar activity were qualitative and aimed to search for periods of active/quiet sun (e.g., Zeller et al., 1986; Dreschhoff and Zeller, 1998) or even for a supernova event (Dreschhoff and Laird, 2006). Later, the method was explored more fully (McCracken et al., 2001a,b), which made it possible to identify large SEP events since 1560 and evaluate their fluence (see Table 1 in McCracken et al., 2001b). Only events with F30 fluence exceeding 109 particles per cm2, which is estimated as a threshold for the nitrate signal, are identified in this way. This analysis leads to a statistical estimate of the occurrence frequency of large SEP events (Figure 24). Thus-obtained statistics of SEP-event occurrence complements directly observed data and that reconstructed from lunar samples, indicating a kind of break in the distribution with the fluence above 1010 protons/cm2. Note that estimates based on lunar-rock samples (arrows in Figure 24) provide only an upper bound for SEP fluence, since they are based on an extreme assumption that the net fluence was produced in a few extreme events. Therefore, extra-strong SEP events with the fluence exceeding 1012 cm–2 are very improbable.
However, the method has a drawback as noted by McCracken et al. (2001b). The measured nitrate effect depends not only on the total fluence F30 but also on the energy spectrum of SEP, namely, high energy particles produce more ionization and thus lead to more nitrate effect than lower-energy particles. Accordingly, the measured nitrate signal can be converted into the SEP fluence only by assuming a fixed energy spectrum of SEP by normalizing all data to, e.g., the event of August 1972 (McCracken et al., 2001b; Shea et al., 2006). However, the actual energy spectrum of SEP remains a free parameter in such a reconstruction (Usoskin et al., 2006b). Moreover, the nitrate record is only an indirect proxy of strong SEP events whose efficiency is not 100% (Palmer et al., 2001).
This work is licensed under a Creative Commons License.