The exchanges of NOx between snow and air have significant impact on the atmospheric components and photochemical processes in the overlying boundary layer. Such exchanges increase the oxidizing capacity of the atmosphere and may have a crucial impact on the air signals that are retrieved from ice cores. In the recent years, sunlit snow and ice have been demonstrated to be important NOx sources in the polar atmospheric boundary layer. This paper makes a thorough review on the release of NOx from snow and ice, including field observations and experimental evidences, release mechanisms and influential parameters that affect such a release process, polar NOx concentrations and fluxes, and environmental impacts of the chemical processes of NOx in the polar atmospheric boundary layer. In the Tibetan Plateau, the released NOx observed recently in the sunlit snow/ice-cover is 1-order magnitude more than that in polar regions, but further scientific research is still needed to reveal its impact on the atmospheric oxidizing capacity.
NOx exchange between snow and air ; polar area ; atmospheric oxidizing capacity ; Tibetan Plateau
NOx (NO+NO2 ) are the main nitrogen containing compounds in the atmosphere. The sources of NOx include fuel combustion, soil emission, biomass combustion, lightning, NH3 oxidation, and stratosphere, etc. [ Tang, 1990 ]. The level of NOx is of ppt (10- 12 ) in the clean atmosphere above ocean and polar region and several to dozens ppb (10- 9 ) in ambient air in cities. It can reach hundreds of ppb in polluted regions. NOx are important reactive gases in the atmosphere in that they determine the ozone concentration in the troposphere, and influence the oxidizing capacity of atmosphere [ Thompson, 1992 ; Atkinson, 2000 ], they thus indirectly affect the lifetime of some greenhouse gases, such as CH4 , HCFCs, etc.
Active gases can be produced by the photolysis of compounds in the snowpack and then released to the boundary layer atmosphere. It has been found that NOx and other gases are released from sunlit snow-pack in Antarctic and Arctic region. These gases had a significant influence on the chemical processes of atmosphere in polar boundary layer and on the interpretation of the historical atmospheric compositions derived from ice core [ Honrath et al., 1999 ; Sumner et al., 1999; Davis et al., 2001 ]. This has brought scientific awareness on the importance of photochemical process on polar snow/ice surface [ Wolff and Shepson, 2003 ; Bottenheim et al ., 2002 ; Dominé and Shepson, 2002 ; Davis et al ., 2004b ; Grannas et al ., 2007 ]. The scientific purposes of the ALERT 2000 and SUMMIT 2000 campaigns conducted respectively at Canadian Alert and Greenland Summit include the snow/air interactions and photochemistry both within and above the snowpack surface [ Bottenheim et al., 2002 ]. After discovering considerable release of NOx from snowpack in Antarctic Plateau in 1998, the goal of ISCAT 2000 campaign was shifted to study NOx /HOx chemistry [Davis et al., 2004a].
In this paper, we review the progress of researches on NOx sources from snow and ice.
At Greenland Summit, relatively high NOx concentration was found in snowpack interstitial air, the concentration is 3–10 times higher than that in the surrounding air. It is also higher than the ambient NOy concentrations [ Honrath et al., 1999 ]. It was proposed that NOx was produced by the photolysis of NO3− in the snowpack, then released to the snowpack interstitial air and subsequently to the atmospheric boundary layer. The vertical profile of NOx measured at Summit showed that NOx concentrations were much higher in the snowpack interstitial air than that in the overlying atmosphere [ Dominé and Shepson, 2002 ]. The diurnal change of NOy concentrations at Antarctic Neumayer station is resulted from the exchanging process in snow-air interface [ Weller et al., 1999 ]. Jones et al.  reported that NO and NO2 concentrations in the Antarctic snow interstitial air were 15 ppt and 32 ppt, respectively, higher than that in ambient air. High concentration NO was limited in a shallow atmospheric layer 20–50 m above the ground, and the NO concentration decreased to background level quickly outside the layer [ Helmig et al., 2008a ]. The highest NOx concentrations were found in the air close to the ground in the Antarctic inland, but no evident NOx concentration gradient occurred at the Antarctic seaside. These observations are consistent with the conclusions: the main sources of NOx in polar atmospheric boundary layer are mainly from the Antarctic snowpack [ Davis et al., 2008 ]. All above observations and the results of snow/air flux measurements, field observations and laboratory experiments confirm that NOx released from snow is the main source of NOx in polar boundary atmosphere.
It has been demonstrated that nitrate solution could produce NO2 , NO2− and OH under 300–350 nm light [ Wagner et al., 1980 ], while NO2− could produce NO under 295–410 nm light and NO had been detected in the air bubble of sunlit seawater [ Zafirou et al., 1980 ]. When continuously exposed to 300 nm light, the sub-millimeter-thick ice doped with NO3− can release NO2 [ Dubowski et al., 2001 ]. Photolysis experiments using natural snowfall showed that higher concentrations of NO3− containing in snow could produce higher concentrations of NO, NO2 and nitrous acid (HONO) [ Dibb et al ., 2002 ; Honrath et al ., 2002 ]. The results of laboratory experiments on photochemical reactions in snow/ice further confirmed the quick release of NOx from NO3− containing snow/ice [ Dubowski et al ., 2001 ; Boxe et al ., 2003 ; 2005]. These studies indicate that both natural and synthetic snow containing NO3− , if exposed to light, can produce and release NO and NO2 quickly.
NO3− has strong absorption on the ultraviolet-band light (UV; 260–330 nm) [ Cotter et al., 2003 ], therefore, the concentrations of NOx above ice surface is usually significantly correlated with the change of solar ultraviolet B-band (UVB; 275–320 nm) radiation [ Honrath et al., 2000 ]. At Neumayer station in the Antarctic, NO concentration during the polar night was nearly zero, NO began to increase in October, and reached the maximum in early December. There are relatively high NO concentration in summer and autumn. The occurrence of annual maximum of NO was strongly correlated with the maxima of UVB radiation, but not with the ultraviolet A-band (UVA; 320–400 nm) radiation [ Weller et al., 2002 ]. This suggests NO is released from the photolysis of snow NO3− rather than the photo-decomposition of NO2 in the atmosphere. In 1998, the results of Polar Sunrise Experiment at Alert showed that NOx concentration was very sensitive to solar radiation, and NOx concentration increased as the increasing solar radiation [ Ridley et al., 2000 ]. Experimental results in the Antarctic showed that the maximum release rates up to 1.1×106 molec cm–2 s–1 for NO and 2.1×106 molec cm–2 s–1 for NO2 could be achieved from sunlit snowpack [ Jones et al., 2000 ].
Cotter et al.  proposed a possible mechanism for the photolysis of snow nitrate (NO3− ) into NO/NO2 :
The main product from NO3− photolysis is gaseous NO2 , which has been confirmed by many field observations [ Jones et al ., 2000 ; Dibb et al ., 2002 ; Jacobi and Hilker, 2007 ]. However, not all NO2 can be released from the snow; only those in snow surface can be directly released into the air and the rest will be further photo-dissociated into NO [ Dubowski et al ., 2001 ; Boxe et al ., 2005 ].
Factors controlling the release of NOx from snow/ice surface are complex. The physical properties of snow, such as temperature, pH and ion content, will affect the chemical reaction and physical diffusion process, and also determine whether the products can be released [ Dominé et al ., 2008 ; Pinzer et al ., 2010 ]. The metamorphism of surface snow will lead to reduction of surface area and thus the adsorbed trace gases will be released [Cabanes et al., 2002]. There are data showing that the formation of nitrous acid (HONO) depends on the pH of the quasi-liquid layer. When pH≥5, HONO production is zero [ Jacobi and Hilker, 2007 ]. The data from field measurements [ Beine et al., 2003 ; 2006 ; Amoroso et al., 2005 ] support the same conclusion that no HONO is generated from alkaline snow. Snow NO3− is likely in the form of salt and trapped in the ice, therefore the sunlit products are not easily released [ Beine et al, 2006 ]. In addition to the wavelength and NO3− concentration, the temperature and the depth of light penetrating into the ice will affect the photolysis reaction. The microbial activity in snow can also lead to NOx emissions [ Amoroso et al., 2010 ], while the ice structure and porosity will determine the ability of the products escaping into the air [ Albert et al., 2002 ]. Furthermore, products generated from NO3− photolysis will react with other chemical components existing in ice and snow. This reaction not only changed the chemical compositions of ice and snow, but also affected the concentration of NOx in the atmosphere. Once released into the atmosphere, NOx will react with other substances, especially with the OH radical to form HNO3 , which can return to the ice surface by dry or wet deposition and continue to participate in the photochemical cycles [ Davis et al., 2008 ].
At Amundsen Scott station in Antarctica in 2000, the average concentrations of NO were (102±69) ppt during November 15–30, (95±103) ppt during December 1–15, and (113±102) ppt during December 15–30, slightly lower than those observed in the same period in 1998 [Davis et al., 2004]. During ANTCI 2003 campaign, the NO concentration of 200–1,000 ppt were observed in surface air and 0–500 ppt by aircraft observations over Antarctic Plateau, and 3–35 ppt by tethering balloon along the Antarctic coast [ Davis et al., 2008 ]. At Antarctic coastal Neumayer station, the observed NOy mixing ratios were (46±29) ppt during the whole year and (58±35) ppt from February to May. The air with NOy mixing ratio greater than 100 ppt usually came from the inland of Antarctic Plateau [ Weller et al., 2002 ]. At Arctic Alert regions, NOx concentrations were close to those measured at Summit, but with less NO concentration [ Beine et al., 2002 ]. The concentrations of NOx were less than 30 ppt at Alert region [ Muthuramu et al., 1994 ]. In Antarctic Summit, NO and NOx mixing ratios showed less fluctuations during the year and between the years [ Yang et al., 2002 ].
The snow/gas fluxes at the polar areas are also measured. At Neumayer station, NOx emissions fluxes were 0–3×108 molec cm–2 s–1 , with an average of 1.3×108 molec cm–2 s–1 in the summer of 1999 [ Jones et al., 2001 ]. In the summer of 2000, the NOx and HONO fluxes over the snow at Antarctic Summit were 2.52×108 molec cm–2 s–1 and 4.64×107 molec cm–2 s–1 respectively [ Honrath et al., 2002 ]. Obtained through eddy covariance methods, the average NO and NOx fluxes over the snow at South Pole in 2000 were (2.6±0.3)×108 and (3.9±0.4)×108 molec cm–2 s–1 respectively [ Oncley et al., 2004 ]. It seems that there were few differences of NOx fluxes between Antarctic and Arctic snow. The positive NOx flux between snow/air indicates that NOx released from snow is an important source in the atmosphere of polar boundary layer.
The photochemical productions of NOx, formaldehyde, and bromine compounds in snowpack can further take photolysis reactions and then trigger the production of HOx free radicals, which can strengthen the atmospheric oxidation [ Sumner and Shepson, 1999 ]. In 1998, the median NO concentration of 225 ppt was observed at the South Pole, which enhanced the concentration of OH radical up to 106 molec cm- 3 [ Davis et al., 2004a ]. OH radicals are produced through the photolysis of HONO released from polar snowpack, and they have an important contribution to the atmospheric oxidation [ Honrath et al ., 2002 ; Beine et al ., 2006 ; Amoroso et al ., 2005 ; Davis et al ., 2004a ; Zhou et al ., 2001 ]. Higher levels of OH (2.5×106 –3.5×106 molec cm- 3 ) and HO2 + RO2 (10100 times of the OH concentration) were observed at the South Pole than the levels observed in the Antarctic marine boundary layer. These observations are consistent with model results that the high level of NOx was important for the high level of OH and HO2 + RO2 [ Mauldin et al ., 2001 ; Mauldin et al ., 2004 ]. Strong atmospheric oxidation can influence the chemical forms and environmental behavior of polar Hg and other substances [ Schroeder et al., 1998 ].
Since high concentrations of NOx released from snowpack are usually limited to a narrow layer of air close to the snow surface [ Helmig et al., 2008a ; 2008b ], this layer should have a strong photochemical oxidation environment. This has been confirmed by computer modeling. Wang et al.  combined the field observations with chemical transport models to evaluate the effect of NOx emissions from the Antarctic snow cover on the photochemical process, and the results showed that the surface of the Antarctic continent is covered with a shallow canopy of high OH concentrations due to snow NOx emissions. The OH radical concentration is greater than 3×106 molec cm- 3 from the surface to 50–150 m height of canopy. The results of OH, NO, and photochemical flux measured in the Antarctic Plateau indicate that the Antarctic has a highly photochemically active oxidizing environment. Excluding the effects of transport and stratospheric ozone depletion, ground and sounding observations also showed the photochemical production of O3 in the Antarctic [ Crawford et al., 2001 ]. More UV radiation on the snow surface due to stratospheric ozone depletion will lead to more NOx emissions from snow, which may be one of the reasons for the increasing tropospheric ozone over the past 30 years in the Antarctic [ Jones and Wolff, 2003 ].
The lifetime of NOx in snowpack determines the chemical composition of the atmosphere near the ground and the concentrations of nitrogen or other components in snow surface and glacier ice. When the photochemically generated NOx is released from the snowpack and the closest boundary layer into the troposphere, it involves in the competition process of vertical mixing and the reaction between NOx and HOx to form HNO3 and HO2 NO2 . The latter two substances can quickly return to the snow by deposition [ Honrath et al ., 2002 ; Oncley et al ., 2004 ; Munger et al ., 1999 ; Cohen et al ., 2007 ]. Deposited HNO3 and HO2NO2 in the snow surface can also be re-decomposed by photolysis. These reactions forms a loop, but not a closed loop, as NOx can be transported to somewhere else by advection, resulting in loss of nitrogen in the snow [ Honrath et al ., 2002 ; Weller et al ., 2002 ].
Therefore, the photochemical process on the surface of polar snow/ice influences not only the atmospheric chemistry of the polar boundary layer, but also the retrieval of atmospheric compositions recorded in the ice cores. Strong photochemical oxidation process must be considered in the interpretation of the atmospheric composition data derived from snow and ice. The variations of NO3− concentration in ice core are used to analyze the past climatic and environmental changes and the impact by human activities. NO3− concentration in ice cores from the Tibetan Plateau reflected the changes in land use, biomass burning, and industrial processes [ Thompson et al., 2000 ]. The high-resolution NO3− concentrations in ice cores from European Alps were used to study the impact of meso-tropospheric NO3− by the increasing NO emissions and to deduce the natural emission intensity of NO in the pre-industrial era [ Preunkert et al., 2003 ]. These studies and analysis did not consider the photochemical processes on ice surface, such as the photolysis of snow nitrate (NO3− ) and the release of NOx . Clearly, the photochemical reaction is likely to affect the concentration of NO3− in ice core; therefore it will affect the interpretation and analysis of the impact of human activities on the change of climate and environment. Other substances in ice cores like H2 O2 , HCHO and CH4 also can be affected. Currently, some related research work is underway [ Frey et al ., 2009 ; Anastasio and Chu, 2009 ].
In fact, the release of NOx in the photolysis experiments of natural snow in winter was observed not only in the polar regions but also at remote areas in the mid-latitudes [ Honrath et al., 2000 ]. Furthermore, the photochemical production of CO from new snow was also observed at Austria’s Sonnblick Hill with an altitude of 3,106 m above sea level [ Haan et al., 2001 ]. Therefore, it is a common phenomenon that NOx or other trace substances are released from snow/ice under solar radiation, but the importance of the release intensity is different in different regions. It often can be negligible in the area with strong anthropogenic emissions, but it also can be very important in very clean regions, where the emission contribution and its potential environmental impact are relatively more important.
Tibetan Plateau, with an area of 2,500,000 km2 , is a significant snow-covered region in the northern midlatitudes. The solar radiation is very strong because of the clean air, high surface reflection and high elevation (above 4,000 m). The low atmosphere has strong photochemical oxidizing capacity [ Lin et al., 2008 ]. Under the influence of westerlies and Asian monsoon, the chemical compositions of snow ice in the north and south part of the plateau is different, but the concentrations of NO3− at the Tibetan Plateau are higher than those in Arctic and Antarctic snow ice [ Li et al., 2000 ]. Therefore, photochemical reaction in snow surface at the Tibetan Plateau will be stronger than that in the Arctic and Antarctic area, as well as the concentrations of NOx released. The observations at the snow surface in July 1st Glacier [ Lin , 2005 ], Yulong Snow Mountain, and Tianshan No. 1 Glacier at Uriimqi River [ Wang , 2008 ] support this conclusion. In these observations, the variation of NOx concentrations was significantly correlated with the change of UVB radiations. The nighttime NOx levels near snow surface were under the detect limit of the instrument, but the daytime NOx levels were 1-order magnitude more than those in polar regions and higher than those observed at Waliguan GAW global station [ Yu et al., 1997 ]. Because of harsh natural environment and mountain terrain, it’s challenging to carry out the comprehensive and in-depth experimental observations in the Tibetan Plateau. It needs more work to estimate the NOx flux of snow/air exchange and their mechanisms.
Compared with the Alps Mountains, which are surrounded by developed countries in Europe, the impact of human activities on the environment of the Tibetan Plateau is much smaller. For example, a nearly half-century ice core record in the Tibetan Plateau shows that the concentrations of SO42 − and NO3− did not rise and their values fluctuate strongly near the average values since human industrialization [ Yao et al., 2000 ]. However, the NO3− concentration from the ice core in the middle latitudes of Europe Alpine shows a clear upward trend during 1850–1960 [Preunkert et al., 2003]. Thus, observations in the Tibetan Plateau may be able to reveal the broader process of environmental impact. The studies on snow/air exchange and on the photochemical reaction in snow ice surface in the Tibetan Plateau are significant. It will contribute to discover the photochemical reactions, especially the oxidizing capacity and its potential role in global change in the Tibetan Plateau. It will be helpful to use the accurate ice core records to analyze and interpret the climatic and environmental changes by human activities. All these require more in-depth investigation.
Many field observations and experimental evidences show that NOx and other substances can be released from snow/ice under solar radiation in polar areas. The exchanges of NOx between snow and air have a significant impact on the atmospheric components and photochemical processes in the overlying boundary layer. These exchanges increase the oxidizing capacity of the atmosphere and may have a critical impact on the signals of atmospheric components that are retrieved from ice cores. These results are important references for the similar researches in the Tibetan Plateau, which has vast snow cover in the midlatitudes in the Northern Hemisphere. The released NOx observed recently from the sunlit snow/ice-cover in the Tibetan Plateau are 1-order magnitude more than those observed in polar regions.
This work is supported by the Fund of Polar Scientific Research (No. 20080216) of State Ocean Administration, China, and by Chinese Natural Science Foundation (No. 20407001, No. 40701170).