Air Masses and Weather Types : A Useful Tool for Characterizing Precipitation Chemistry and Wet Deposition

This study is an analysis of 344 days with rainfall recorded during five years in a remote regional background EMEP (Cooperative Programme for the Monitoring and Evaluation of the Long Range Transmission of Air Pollutants in Europe) station in Spain. The chemical composition of the rainwater associated with air masses (nine categories) and weather types (26 categories) was characterized. The chemical composition of rainwater was dominated by calcium (Ca) and sulphate (SO4-S), with VWM (Volume Weighted Mean) during the period studied (2002–2006), with 55 μeq/L and 34 μeqS/L, respectively. Calcium, sodium (Na), ammonium (NH4-N) and magnesium (Mg) seem to be dominant components in the neutralization of the rainwater. By applying Pearson correlations, principal component analysis and enrichment factors, it is possible to identify source types for the precipitation constituents. Interannual and intra-annual variability was also been studied. High calcium levels are associated with the frequent intrusions of Saharan dust that occur during the summer, and the maximums of chlorine and sodium in the winter may be due to the greater amount of maritime air recorded during this season. Wet deposition was determined by focusing on nitrogen deposition, registering mean annual values of 155 mgN/m/year (from the NO3-N) and 165 mgN/m/year (from the NH4-N).


INTRODUCTION
Rain acts as a powerful mechanism to remove pollutants from the atmosphere.Precipitation chemistry is the result of a series of in-cloud and below-cloud atmospheric chemical reactions and a complex interaction between microphysical processes and cloud dynamics (Mouli et al., 2005).The importance of studing chemical composition of rainwater focus on several aspects: i) it is a valuable tool that helps us to identify the pollutant sources involved in this composition; ii) it provides information on the transportation and dispersion of pollution; iii) it is involved in problems related to acid deposition, eutrophication, trace metal deposition, biogeochemical cycling, ecosystem health and global climate change and iv) it is a very useful tool for validating model simulations of air pollution (Zunckel et al., 2003;Sanets and Chuduk, 2005;Brimblecombe et al., 2007;Zhang et al., 2007;WMO, 2008;Budhavant et al., 2009;Galy-Lacaux et al., 2009;Yi et al., 2010).
Due to its crucial importance, precipitation composition has been and is systematically studied all over the world (e.g.Tanner, 1999;Herut et al., 2000;Okai et al., 2002;Zhang et al., 2007;Celle-Jeaton et al., 2009;Galy-Lacaux et al., 2009;Huang et al., 2009;Beem et al., 2010;Calvo et al., 2010;Das et al., 2010a, b;Osada et al., 2011).Aerosols and gases released into the atmosphere can be transported over long distance from their source, and can be removed by dry or wet deposition (e.g.Fraile et al., 2006;Shen et al., 2011).Background levels, established for remote areas far from the direct impact of anthropogenic sources, provide relevant information mainly because it allows us to determine the extent of anthropogenic pollution as well as to address the potential influence of long-range transport to rural areas.Networks of wet deposition are very useful to establish appropriate thresholds (Das et al., 2010a) for policy decisions.Among the several extensive networks of wet deposition in operation in different places around the world (Japan Environmental Agency (JEA) Network, East Asian Network (EANET), National Atmospheric Deposition Program (NADP-in North America), etc.), Europe has the European Monitoring and Evaluation Program (EMEP).The EMEP network consists of 150 stations distributed in 35 countries throughout the whole of the European continent.Reports with information on emissions, concentrations and depositions of air pollutants, transboundary fluxes, exceedances to critical loads and threshold levels, etc. are frequently published (EMEP, 2012).These reports show that transboundary transport is responsible for a significant fraction of the pollution in European cities, as well as in rural areas (e.g., EMEP-WMO, 1999).Thus, for example, Krzyzanowski, (1999) pointed out that as much as 40-60% of PM 10 levels may be attributable to long-range transport in many populated areas, mainly in places where there are no heavily polluting local sources of PM.The study carried out by Kindap et al. (2006) in Istanbul showed that transboundary sources may be responsible for as much as half of the background PM 10 .Hacisalihoglu et al. (1992) in their study near the Black Sea, found that 70% of the mean concentrations of various pollutants were originated from Central and Western Europe.
Ten EMEP stations are currently operative in Spain, distributed all over the country (EMEP, 2012).
Among the different species present in rainwater, the deposition of atmospheric nitrogen (NO X , NH Y ) has an important relevance due to its large influence on terrestrial and aquatic ecosystem.Thus, an excessive input of reactive N could cause important an numerous ecological perturbations (Matson et al., 2002;Yu et al., 2011) such as decrease of plant diversity (Bobbink et al., 2010;Stevens et al., 2010), eutrophication of surface waters (Bencs et al., 2009;Rojas and Venegas, 2009), or changes of greenhouse gas flux (Jiang et al., 2010;Li et al., 2012).
It is important to mention that, at global scale, reactive N has increased from 15 Tg N in 1860 to 187 Tg N in 2005 (Galloway et al., 2008) and, according to recent studies (Nakicenovic and Swart, 2000;Galloway et al., 2004;Lamarque et al., 2005), the increasing trend is considered to be enhanced in the next few decades.Trends of monitored nitrogen in Europe have been studied by several authors (Sopauskiene et al., 2001;Puxbaum et al., 2002;EMEP, 2004a, b;Honová et al., 2004;Fowler et al., 2005), and they conclude that, in many countries, a large part of the nitrogen is emitted by sources outside the countries where the measurement sites were situated, so transboundary fluxes need to be taken into account (Fagerli and Aas, 2008).
Current and future (2030) deposition of reactive nitrogen to land and ocean surfaces has been estimated by Detener et al. (2006) by using several atmospheric chemistry transport models.They concluded that the main factor conditioning future deposition fluxes will be changes in emissions, whereas changes in atmospheric chemistry and climate will have a lower influence.Thus, it is important to identify the main sources of reactive nitrogen in order to control them and enforce current air quality legislaton, avoiding future important problems (e.g., eutrophying and acidifying deposition of the ecosystems, air quality or radiative forcing) (Detener et al., 2005(Detener et al., , 2006;;Shindell et al., 2006;Stevenson et al., 2006).
This study provides an analysis of 344 days with rainfall recorded between 1 January 2002 and 31 December 2006 in a remote regional background EMEP station in Víznar, Granada, Spain.The chemical composition of the rainwater associated with air masses and circulation weather types (CWT) has been established.Furthermore, by applying various statistical techniques it has been possible to identify data relationships and source types for precipitation constituents.Special attention has been focused on nitrogen deposition.These results significantly contribute to the knowledge on rainwater quality in background stations in Europe, and point out the relevance of long range transport on atmospheric pollution.

STUDY ZONE
The measurements were made in Víznar (37°14'N, 03°28'W; 1260 m a.s.l), a rural area located 12.5 km NE of the city of Granada, in the south-east of Spain (Fig. 1).It is situated on one of the mountains that surround the basin of Granada, which form a part of the Sierra de Huetor Natural Park, a mountainous region of medium height with large masses of holm oaks (Quercus ilex) and Portuguese oaks (Quercus faginea).The geological setting embraces the Alpujarride complex, with its carbonate rocks, the Malaguide complex, of a detritic character, and the Subbetic Domain, with an essentially carbonate nature (Rubio et al., 2008).In order to find low emission densities in the vicinity of EMEP sites, there is a general tendency to locate EMEP sites at higher altitudes; criterion which can more easily be met in mountainous areas (Spangl et al., 2007;EC, 2011).
The climate is of the Continental Mediterranean type, characterized by cold winters, with many severe frosts, while the summers are very warm (www.aemet.es).The mean annual temperature is 13°C.The annual precipitation is around 400 mm (Spanish National Institute of Meteorology, AEMET).March is the month with the highest rain rate (444 mm), and July and August with the lowest, with 3.3 and 4.3 mm, respectively.More detailed information about this study zone has been reported by Calvo et al. (2010).

Sampling and Analysis Methodology
Wet deposition samples were collected at the Viznar EMEP station between 1 January 2002 and 31 December 2006, with a total of 344 samples on a daily basis.Integrated samples were collected daily at 0700 UTC and stored in a refrigerator until they were sent, once a week, to the laboratory at the Carlos III Health Institute in Madrid, Spain.Rain samples were analysed in order to obtain the concentration of sodium, magnesium, calcium, chloride, potassium (all in μeq/L), sulphate (μeqS/L), nitrate (μeqN/L), ammonium (μeqN/L), conductivity (μS/cm) and pH (EMEP, 2012).
Potassium, sodium, magnesium and calcium were determined using Atomic Absorption or Emission Spectrophotometer and ammonium concentrations were measured by means of the Spectrophotometric Indophenol method.Ionic chromatography was used in order to determinate sulphate, nitrate and chloride concentrations.Selection criteria, sampling procedures, analysis and data quality controls are pre-established (EMEP, 2000(EMEP, , 2001) ) as Viznar is an EMEP station.

Volumetric Mean
The volumetric mean (VWM) -in μeq/L-and the standard deviation of the volumetric mean (VWSD) have been calculated.The following equations have been used to calculate these two parameters (Galloway and Gaudry, 1984): where C i represents the concentration of a specific element in our sample (in μeq/L except for the pH and conductivity -μS/cm), P i is the amount of precipitation recorded for each event (in millimetres) and N is the number of samples.
Using the VWM, we calculated the concentration of an ion in a given period (one month, one year, etc.), as a result of which we take into account the influence of the amount of rainfall on the concentration of ions (Staelens et al., 2005).

Enrichment Factors
Enrichment factors (EnF) have been calculated, both for the seawater and for the crustal material (Zhang et al., 2007;Huang et al., 2008), according to the following equation: where (C x /C r ) sample is the quotient of the concentration of an element (x) with respect to a reference element (r) in the sample, and (C x /C r ) crustal or seawater is the same quotient considering the concentrations in the crustal material or seawater.
In our study, we used sodium as a reference element of marine origin, and calcium as a reference element of crustal origin.Following Jordan et al. (2003), Mg 2+ vs. Na + have been plotted and sizing by Ca 2+ (Fig. 2) As indicated by Jordan et al. (2003), it is possible to determine which species to use as the reference sea-salt species by comparing the measured ratio of Mg 2+ /Na + to the equivalence ratio of 0.227 found in bulk seawater (Wilson, 1975;Keene et al., 1986).A ratio in excess of 0.227 indicates a crustal Mg 2+ influence, while a ratio less than this suggests a crustal Na + component.In this study, Mg 2+ is clearly enhanced, altering the slope significantly from what one would expect for sea salt.Regarding Ca 2+ , high values of Mg 2+ are linked to high values of Ca 2+ , indicating a similar temporal variation.The smaller points (lower Ca 2+ concentration) are closer to Mg 2+ -Na + seawater points.Thus for this study, Na + is used as the reference species for sea salt, since it appears less likely to suffer from major deviations from the marine source.
The quotients of the concentrations of the different elements in seawater (C x /C Na ) and crustal material (C x /C Ca ) were obtained based on the studies of Keene et al. (1986) and Taylor (1964), respectively.
Elements with values higher than 10 are assumed to come from other sources apart from the crustal or marine source, for EnF crustal and EnF seawater respectively (Báez et al., 2007).
It has been estimated that organic acids contribute between 25% and 65% of the rainwater acidity in industrial and urban areas (Yu et al., 1998;Altieri, 2009) and this contribution is higher in remote areas (Galloway et al., 1982, Likens et al., 1987, Andreae et al., 1988;Yu et al., 2009).
Due to the lack of organic acid concentrations, neutralization factors have been calculated only considering SO 4 2− and NO 3 − , following Kulshrestha et al. (1995) where X is the alkaline component (Ca 2+ , NH 4 + , Na + , Mg 2+ ) with all of the ions expressed as μeq/L.

Statistical Procedures
Principal Components Factor Analysis (PCA) was applied in order to explore data relationships and source types for precipitation constituents.This multivariate statistical method allows to reduce a multidimensional system with many correlated variables into a simpler system of uncorrelated variables, called principal components, which explain a high percentage of variance of the original system (Calvo et al.,  2010).By means of the correlation of the variables to the factors, it is possible to interpret the factors obtained and to identify the potential source of variance associated with each factor (Drever, 1982).
Correlation analysis has been applied in order to examine the relationships between the ions analysed.

Trajectories
Five days long back-trajectories for three different altitudes -500 m, 1500 m and 3000 m a.g.l.-calculated with the HYSPLIT model - Draxler and Rolph (2003)-, using the FNL (Final analysis) database, were used in order to study the influence of different air masses on the chemical precipitation in our study zone.The precipitation events were classified into nine groups (Mediterranean, Tropical Maritime, Polar Maritime, Local, Continental, Arctic, Saharan 500 m, Saharan 1500 m and Saharan 3000 m) following the methodology described by Calvo et al. (2008,  2010).Saharan category has been segregated into three altitudes aiming to reveal the influence of Saharan air masses at high levels on rainwater chemical composition (Fig. 3).
A detailed description of the methodology carried out for the air masses classification can be found in Segura et  al. (2012).

Circulation Weather Types
A Circulation Weather Types (CWTs) classification was carried out based on Jenkinson and Collison (1977) and Jones et al. (1993), in order to identify the type of weather associated with a particular synoptic situation.The daily circulation affecting the Iberian Peninsula is described using a set of indices associated with the direction and vorticity of the geostrophic flow.The indices used were the following: southerly flow (SF), westerly flow (WF), total flow (F), southerly shear vorticity (ZS), westerly shear vorticity (ZW) and total shear vorticity (Z).These indices are computed using sea level pressure values (SLP) obtained for the 16 grid points distributed over the Iberian Peninsula.This database is available for the majority of the northern hemisphere in intervals of 5° of latitude by 5° of longitude.The same grid was used as in the study carried out by Trigo and DaCamara (2000).This method allows for a maximum of 26 different CWTs.Two of them are the so-called "nondirectionals": "Pure anticyclonic" (A) and "Pure cyclonic" (C).The next eight weather types are known as "pure", and are characterised by a specific predominant wind component, without emphasising the influence of a high or low, and they are N, S, E, W, NW, SW, SE and NE.
The remaining sixteen weather types are the so-called "hybrid" types, a result of the combination between one or two cyclic types with the eight predominant wind types (AN, ANW, AW, ASW, AS, ASE, AE, ANE, CN, CNW, CW, CSW, CS, CSE, CE and CNE).Extended information on this classification may be found in Trigo and DaCamara (2000) and Castro et al. (2010).

Rainwater Chemical Composition and Wet Deposition
The entire database enables to estimate the mean chemical composition of precipitations and the corresponding wet deposition over the 2002-2006 period.Table 1 summarizes the annual VWM means and associated standard deviations.
To better characterize dominant meteorological influences and airmasses pathways, the database of rainwater chemical composition has been also analysed through both classifications; Fig. 4 and Fig. 5 provide these clustered values of the VWM.

Nitrogen Contribution
In order to evaluate nitrogen deposition, the determination of inorganic nitrogen (NO 3 − -N and NH 4 + -N) has been considered as an effective approach and it has been widely adopted in America and Europe (Holland et al., 2005; Yu  et al., 2011).In Viznar, the annual VWM concentrations of nitrate and ammonium are 22.5 and 24.8 μeqN/L respectively.After calcium and sulphate, they stand for the third most important contribution to the rainwater chemical composition.

a) Nitrate
Nitrate is one of the nitrogenous contribution to the chemical composition of precipitation.The HNO 3 (gaseous) which results from the oxidation of NO x is water soluble, meaning it is washed away by precipitation, constituting one of the most important sources of NO 3 − in rainwater (Kumar, 1986).In the case of nitrate, it is important to consider the variations in the efficiency of conversion from NO 2 to gaseous HNO 3 ; in summer, this conversion occurs rapidly, while in the winter it is slower (Van Egmond and  Kesseboom, 1985; Ruijgrok et al., 1992).Also, a higher concentration of NO 3 − in the summerly precipitations (Table 2) may be explained by the increase in the production of HNO 3 from gaseous precursors during photochemical activity, and by a displacement of the balance from the particulate phase (NH 4 NO 3 ) towards the gaseous phase (NH 3 and HNO 3 ) (HNO 3 + NH 3 ⇌ NH 4 NO 3 ) (Stelson and  Seinfeld, 1982; Gupta et al., 2003).
With regard to nitrate in link with synoptic situation and meteorological conditions, the highest VWM values were recorded in pure circulation of the N, NE, E and SE between 28 and 35 μeq/L NO 3 − (with mainly easterly winds).None of the hybrid cyclonic weather types exceed 24 μeq/L.In the hybrid anticyclonic weather types, the highest values were recorded for the AN, ANE, AE and ANW types, with values of between 40 and 94 μeq/L NO 3 − .The Saharan and Local categories stands for the highest VWM concentrations which means that both long range transport and local sources influence the nitrate chemistry.The high VWM mean concentrations in Saharan origin airmasses are discussed in the acid contribution part.
The Mediterranean category includes the highest concentrations of NO 3 − among the three maritime categories.The Mediterranean zone is characterised by a complex meteorology that favours the aging of the contaminated air masses in the basin (Millán et al., 1997) and induces high levels of atmospheric particles (Artiñano et al., 2001;  Querol et al., 2001; Lyamani et al., 2006).As a result, the Mediterranean basin may represent a temporary reservoir  (for anthropogenic aerosols) and an additional source (potential seasalt contribution) of atmospheric aerosols for the study area (McGovern et al., 1999).

b) Ammonium
The second contribution to nitrogeneous rainwater composition is of ammonium.Its presence in precipitations results of the condensation of aerosols containing ammonium and of the incorporation of gaseous ammonia in cloud droplets.Major sources of ammonia are known to be natural or fertilized soils, excrements of human and animals, and wood burnings.Maximum ammonia concentration were registered during summer: 36, 30, 22, 17 μeqN/L in summer, spring, winter and autumn, respectively), coinciding with the minimum precipitation registered (Table 2).
Ammonium has the highest values in pure circulation for the N, E and SE (with easterly winds) of between 30 and 48 μeq/L NH 4 + .In the pure anticiclonic category presents the second higher value with 51 μeq/L NH 4 + .In the hybrid cyclonic weather types 30 μeq/L NH 4 + was not exceeded, while in the hybrid anticyclonic weather types there are values in excess of 50 μeq/L NH 4 + for AN and ANE.These results show that the hybrid types with cyclonic circulation are associated with lower volumetric concentrations, due to the significant dispersion of contaminants that occurs.
High VWM concentrations of nitrogeneous compounds are found in airmasses of Saharan, continental and local types.The events of arrivals of European air masses have been explored by authors such as Carratalá and Bellot (1998), Alonso et al. (2000), Gangoiti et al. (2002) or Viana et al. (2003).It is complicated to establish which part of the    pollution is attributable to long-distance transportation from Europe and/or from local or regional emissions (Viana et  al., 2003; Escudero et al., 2007).
The nitrate and ammonium ions increase their concentrations with low vorticities from the W and S (i.e. they are negatively correlated with ZW and ZS) (Table 4). of water precipitated plays an important role in the concentrations of the the different ions studied.It can't be forgotten that previous studies have observed that most of the scavenging occurs in the first 2 mm of rain registered -between 40% and 80% of the wet-only deposition of the major ions (Alastuey et al., 2001).
It can be observed an important variation of VWM with time, with an opposite trend between precipitation heights and nitrogenous aerosol concentration (Fig. 6) (EMEP, 2012).This means that wash out is efficient to clean up the air of aerosols.But this contribution of nitrogenous aerosol compounds through wash out in total nitrogenous concentrations in precipitations does not seem to affect the variability of the VWM.Thus there are no seasonal constraints from the nitrogenous aerosol component on the full signal of nitrogenous concentrations in precipitations.
Thus this nitrogenous aerosol component may be diluted in the whole signal of direct precipitation of nitrogenous compounds.
If we focuss on the ratio NH 4 + /NO 3 − , we can see that it has increased during the last year of study due mainly to a decrease in NO 3 − deposition and the increase in NH 4 + .Yearly mean NH 4 + /NO 3 − ratio varied from 0.89 to 1.45, with an average value of 1.1 in the observation period.If we focus on seasonal ratios, values of 1.1, 0.7, 1.0 and 1.2 were registered for spring, summer, autumn and winter.These values show a similar contribution of NH 4 + and NO 3 − for inorganic-N deposition, i.e., agricultural and industrial activities.
The slightly higher ratios in winter and spring could be related to the NH 3 volatilization induced by application of N fertilizer and/or for the soil freeze/thaw cycle that could promote NH 3 volatilization (Edwards and Killham, 1986;  Yu et al., 2011)

Acidic Contribution
No acid rain problems were detected in Víznar (the pH was always higher than 5.6), as the pH of the rainwater collected varied between 5.7 and 7.5 (the average value for the whole of the study interval was 6.5 ± 0.4).Acidity in rainwater may be neutralized by mineral dust , high loaded in carbonates, or by ammonium.Potential acidity of precipitation is often linked to the presence of organic acids or mineral ones, such as sulphuric and nitric acids.The organic part of acidity cannot be discussed here as formate or acetate were not analysed here.
A labelled "acidity factor" has been found when PCA was applied (Table 5), This factor explains 11.5% of the total variance of the investigated data set and include high loadings negatively correlated for pH and H + , as expected.
The neutralization of rainwater is sustained by a good correlation (Table 4) observed between the SO 4 2− -S and NO 3 − -N with the Ca 2+ , Mg 2+ and K + with r > 0.79, as well as with the Na 4 + -N with r > 0.53.These important correlations may be due to the atmospheric chemical reactions between these elements (absorption of acidic elements such as sulphuric acid and nitric acid on the aerosols, and their subsequent reaction with rich alkaline components in the carbonates present in the particulate material) (Tu et al.,  2005; Zhang et al., 2007).As it can be observed in Fig. 7, the calcium, sodium, ammonium and magnesium ions seem to be dominant in the neutralisation of the rainwater (Ca 2+ has the highest NF (1.1 ± 0.8), followed by NH 4 + -N (0.4 ± 0.3), Na + (0.4 ± 0.3) and Mg 2+ (0.3 ± 0.2).The potassium has the lowest NF, with a value of 0.07 ± 0.04).
No significant correlation were registered between the H + and SO 4 2− -S and NO 3 − -N (r = -0.20 and -0.26, respectively) due to the important role that Ca 2+ plays in the neutralisation process (Topçu et al., 2002), as can be seen when we observed that Ca 2+ has the highest NF.Also, in the atmosphere ammonium-mainly associated with farming and biomass burning-mainly combines with sulphate and nitrate to form ammonium sulphate and ammonium nitrate, respectively (Meng and Seinfeld, 1994; Redington and  Derwent, 2002; Pathak et al., 2009).
High pH levels, as well as high concentrations of cations (mainly Ca 2+ ) and marine ions characterise the chemical composition of the rainwater associated with incursions of Saharan dust (Saharan (n = 62) and Saharan 1500 m (n = 44) (Savoie et al., 1992; Prospero et al., 1995; Ávila and  Alarcón, 1999; Kandler et al., 2007).The events included in Saharan 3000 m (n = 23) category have a typically  Continental and Mediterranean origin at low levels (500 m and 1500 m).If we compare the precipitation events included in the Saharan 3000 m type with those classified as Continental or Mediterranean, the Saharan 3000 m category has higher volumetric means for almost all the elements (except for K + , H + , Ca 2+ and the pH).However, applying the non-parametric Mann-Whitney test reveals that there are no significant differences for any of the ions studied in the precipitation between the Saharan 3000 m category and the Continental and Mediterranean categories, which indicates that the Saharan 3000 m air mass does not seem to have a marked influence on the chemical composition of the precipitation.
When analysing the characteristics of the days included in each category, it is necessary to take into account that some specific situations may arise that cause their characteristics to vary from those expected due to their origin.For example, some days characterised as Saharan do not have the typical chemical characteristics described for 'red' African rains with high alkalinity and a high concentration of cations (Loye-Pilot et al., 1986, 1990; Rodà et al., 1993; Avila et  al., 1997).As Avila and Alarcón (1999) discuss, this may be due to the fact that the rains from Africa may only have these identifying characteristics if they cross through a cloud of dust on their way over the African continent at the correct height.The elevation of dust is more frequent under certain humidity and temperature conditions that occur especially in the spring and autumn (Morales, 1979).Also, some days included in the Polar Maritime category have typical characteristics of Saharan events (basic pH and a high concentration of cations and sulphate).In this case, we interpret that this air mass, despite being defined as maritime as a result of coming from the Atlantic, could be influenced by clouds of African dust moving towards the west over the Atlantic, meaning it would become enriched by African materials.The transport of African dust over the north Atlantic is extensively described in the literature (Prospero  and Nees, 1986; Duce et al., 1991; Prospero, 1996).
Sulphate has the highest VWM values in pure circulation of the N, NE and NW between 35 and 54 μeqS/L (with northerly winds).In the hybrid cyclonic weather categories, the highest values of nearly 39 μeqS/L are with CNE and CS types.In the hybrid anticyclonic weather categories, there are values of around 63 μeqS/L for AN, ANE and ANW.A good correlation was seen between the nitrate and sulphate (r = 0.83) (Table 4), which is attributable to the similar chemical behaviour they have in the precipitation, as well as to the co-emission of their precursors, SO 2 and NO x (mainly due to the burning of fossil fuels, and also to the burning of biomass) (Zhang et al., 2007; Huang et al., 2008).These two elements have an important anthropogenic contribution, as can be seen when we observed EnF crustal (Table 6) with values of 31 and 200 for sulphate and nitrate, respectively.

Marine and Terrigeneous Contributions
The annual VWM concentrations of sodium an chlorine are respectively of 18.7 μeq/L and 24.4 μeq/L (Table 1).The Cl − is very well correlated with the Na + (r = 0.75) (Table 4), suggesting a marine origin for both ions.This fact is confirmed when the principal component analysis is carried out, where both elements are grouped in a factor explaining 6.3% of the variance (Table 5).When enrichment factors are analysed, it can be seen that the Cl − has a clearly marine origin (EnF crustal = 123.37 and EnF seawater = 1.12, by using Na + as reference element) (Table 6).
The Cl − and the Na + present a similar behaviour, with significant joint concentrations in pure circulations of the SW and NW.It was seen that chlorine and sodium jointly present the highest volumetric concentrations in the hybrid weather types CW and AW, with mean VWM values in excess of 30 μeq/L.That is, if the wind component is from the west both with an anticyclone and a cyclone, then there are high chlorine and sodium values.The cyclonic circulation (C) which occurs when there is a stagnant low from the NW zonal flow, and which is usually reinforced by a low at high altitude (500 hPa), favours high concentrations of chlorine and sodium.This fact is corroborated when the Pearson correlation coefficient is observed.Thus, chlorine and sodium are positively correlated with the WF (i.e.high WF values favour the arrival of marine ions) (Table 4).
Calcium is the most abundant species in the Viznar rainwater (annual VWM of 54.8 ± 74.4 μeq/L).The cations Ca 2+ , Mg 2+ and K + are well correlated, with values of r Mg-Ca = 0.94, r K-Ca = 0.55 and r K-Mg = 0.64, suggesting a crustal origin.However, Mg 2+ and K + EnFs (Table 6) seawater with values lower than 10, reveal the existence of a certain influence from the marine source in the concentration of these two elements.The calcium has the highest volumetric concentrations in pure circulations of the N, NE and S, of between 70 and 84 μeq/L.On analysing the hybrid cyclonic categories, we find that the highest values are reached in hybrid weather types: 102 μeq/L for CSE and 77 μeq/L for CSW.However, in the same way as the magnesium, the situation changes in the presence of hybrid anticyclonic weather types (AN, ANE, AE, ANW and AS), as here we find higher concentrations (between 76 and 183 μeq/L).
The magnesium has the highest values in pure circulation from the N and E between 20 and 17 μeq/L, respectively.In the hybrid weather types, the values are between 7 (for CW) and 19 μeq/L (for CSE).However, the situation changes when there are hybrid anticyclonic types, as the concentrations are between 21 and 58 μeq/L for AN, ANE, AE and ANW.Here we can clearly observe that the anticyclone inhibits dispersion.
In the case of the calcium and the magnesium, the South component appears in CSE, CSW and AS, which does not appear for any other ion.The Mann-Whitney test shows that there are no significant differences between these weather types and the Saharan air masses (500, 1500 and 3000 m) for any of the ions that were studied (only for H + when associating the CSW with the Saharan 3000 m category).
These crustal components (Ca 2+ , Mg 2+ and K + ), together with the antropogenic ones (SO 4 2− , NO 3 − , NH 4 + ) and conductivity, appear positively correlated in component one of the PCA, explaining 38.8% of the total variance (Table 5).A "pollution factor" label could be assigned to this factor as this ion grouping may be interpreted as the dominant precipitation chemistry type in this region (Avila and  Alarcón, 1999; Calvo et al., 2010).Furthermore, as result of this PCA analyses, a second factor, explaining 14.7% of the variance, contains the parameters associated with the vorticity (Z, ZW and ZS) and another factor groups together the three parameters associated with the flow of winds (F, WF and SF), explaining 9.4% of the total variance.
The total shear vorticity (Z) is related with the wind vorticity.High values are characteristic of cyclones and anticyclones, and the lowest for pure circulations of the N, S, W, SW, E, etc.The influence in Z comes both from southerly shear vorticity, ZS, (r = 0.739) and westerly shear vorticity, ZW, (r = 0.902).The total shear vorticity Z is negatively correlated with all of the ions except Ca 2+ , Cl − and H + .

Wet Deposition
The wet deposition has been calculated (mg/m 2 /year) for each of the years included in the study by multiplying the VWM concentrations by the annual rainfall amount.The concentrations for days with missing precipitation data have consequently been assumed to be equal to the weighted average of the period (EMEP, 2005).On representing the wet deposition recorded for each element analysed during the five years of the study (Fig. 8) we see that the highest deposited amounts correspond to Ca 2+ , Cl -N (with the average of 155 mgN/m 2 /year) and from 206 mgN/m 2 /year to 378 mgN/m 2 /year for total inorganic N (with the average of 320 mgN/m 2 /year).
In a Romanian site in the forested clean area of Retezeat Mountains gives a deposition of 260 and 330 mgN/m 2 /year for 2000-2002 for N-NO 3 − and N-NH 4 + respectively (Bytnerowicz et al., 2005), similar level at Pop Ivan in the Carpathian Mountains in Ukraine with about 190 and 380 mgN/m 2 /year in 2008 (Oulehle, 2010).
Inorganic-N deposition in Northern Hemisphere, ranging from 40 to 100 mgN/m 2 /year were estimated by Holland et  al. (1999) during the pre-industrial period.In areas very far away from the industrial areas, such as the Qinghai-Tibetan Plateau, atmospheric N deposition as low as 240 mgN/m 2 /year were registered (Jia et al., 2009).However, nowadays, in central Europe, values of wet deposition of total nitrogen above 2000 mgN/m 2 /year have been registered, and in China, values as high as 6225 mgN/m 2 /year have been measured (Lu and Tian, 2007; Yu et al., 2011) ) recorded for each of the elements analysed in the rainwater during the five years of the study.NTN) (a 200-station, rural, wet-only deposition monitoring network throughout the continental United States, Alaska, and Puerto Rico) (Yu et al., 2011).The values registered in Viznar are in the low ranges in the EMEP register.

Interannual and Intra-Annual Variability
A total of 344 rain days were recorded between 1 January 2002 and 31 December 2006December , with 82 events in 2002December , 88 in 2003December , 68 in 2004December , 46 in 2005December and 60 in 2006.With regard to the amount of rain precipitated, we have mean annual values of 660, 578, 519, 276 and 429 mm, for the years 2002, 2003, 2004, 2005 and 2006,  If we focus on the volumetric concentration of ammonium, we see a gradual increase over the years, from 19.69 μeqN/L in 2002 to 32.07 μeqN/L in 2007.This increase was also observed in the particulate phase, where the concentation of (NH 3 + NH 4 + ) increase from 0.4 ± 0.4 μg/m 3 in 2002 to 2.0 ± 1.1 μg/m 3 (Calvo, 2009).Due to the important contribution of local category to NH 4 + deposition, this increase could be related to a change in the soil uses, in the fertilizers used, an increasing fertilization, an increase in the farm activity in the surroundings, etc.Indeed, ammonia emissions at the scale of the Andalucia region increase of about one third between 2002 and 2006 (www.prtr-es.es)(from 5895 to 7458 ton/year).This increase seems to be linked with the intensification of the pig farming and associated emissions (1966 ton/year in 2002 to 3170 ton/year in 2006).Furthermore, an increase in the number of Sahara 500 m events, associated with important concentrations of NH 4 + have been registered during the studied period (28, 31, 35, 48 and 62 events for 2002, 2003, 2004, 2005 and 2006, respectively).This fact could be a possible factor involved in the NH 4 + increased registered.On the other hand, it needs to be taken into account that the qualities of the nitrogen precipitation measurements are in general quite good, as for sulfur the analytical uncertainty is relatively low and the deposition uncertainty is about 10% (Uggerud and Hjellbrekke, 2009).
Focusing on the whole studied period, we can affirm that the chemistry of the rainwater is dominated by Ca 2+ and SO 4 2− (Table 1), with VWM during the studied period (2002)(2003)(2004)(2005)(2006) of 55 μeq/L and 34 μeqS/L, that constitute 28% and 17% of the total, respectively.High percentages (of between 11 and 13%) were found of NH 4 + -N, Cl − and NO 3 − .Na + and Mg 2+ constitute between 9% and 8% of the total volumetric concentration that was recorded, and K + and H + appear as minority elements, with percentages of less than 2% (Table 1).
The maximum concentrations of nitrate were recorded in the summer, as has been previously discussed.Nitrate, as sulphate, showed, applying Mann-Whitney test, significant differences (at significance level of 0.05), between all of the seasons, except between the winter and autumn.
The majority of the maximum average seasonal concentrations of the precipitation events recorded during the study period 2002-2006 were obtained in the summer, except for Cl − and Na + whose maximum is in winter, and H + in the spring.The maximums of chlorine and sodium in the winter may be due to the larger number of arrivals of masses of maritime air recorded during this season.Mann-Whitney test showed significant seasonal differences (at significance level of 0.05), between the winter and autumn and summer and autumn for sodium, and between the winter and spring and spring and autumn for chlorine.They are the ions with the least significant seasonal differences.Mann-Whitney test was applied after observing significant differences for all of the parameters that were studied by applying the Kruskal-Wallis test.
The high volumetric concentration recorded in the summer for calcium is especially striking, with values of 206 μeq/L compared to 35, 56 and 45 μeq/L in the winter, spring and autumn respectively.These high calcium levels are associated with the frequent intrusions of Saharan dust that occur during the summer.Focusing on the seasonal percentages that constitute the average volumetric concentrations recorded for each of the ions studied, we can see that calcium, in the summer, constitutes nearly 45% of the total of the concentration recorded, compared to the 20%, 27% and 28% it constitutes in the winter, spring and autumn respectively.Applying Mann-Whitney test, it can be seen that calcium only has non-significant differences (at significance level of 0.05) between the spring and autumn.The high levels of PM 10 registered during the summer months, with mean concentrations of 32 ± 20 μg/m 3 for summer versus 17 ± 16, 24 ± 17 and 14 ± 11 μg/m 3 for winter, spring and autumn, respectively, during the studied period corroborate these results (Calvo, 2009;EMEP, 2012).Daily mean concentrations of around 200 μg/m 3 were reached during some summer days as a consequence of Saharan intrussions (Calvo, 2009).
It is important to take into account the fact that during the summers, a large number of the precipitation events that took place left behind a small precipitated volume.As a result, we would expect a minimum dilution, and as a result, high volumetric concentrations of the different ions studied.It is therefore logical that the majority of the minimum concentrations of the different parameters being studied were recorded in the autumn, which was the season with the highest mean seasonal volume (Table 2) (Hansen et al.,  1994; Naik et al., 1994; Alastuey, 2001; Zhang et al., 2007).

CONCLUSIONS
This study shows the analysis of 344 days (2002)(2003)(2004)(2005)(2006) of rainwater inorganic chemical composition in a remote regional background EMEP station in Spain.Due to its important impacts, nitrogen contribution has been emphasized.Nitrate and ammonium species (with annual VWM concentrations of 22.5 and 24.8 μeqN/L, respectively) stand for the third most important contribution to the rainwater chemical composition.A similar contribution of both species for inorganic-N deposition (with a yearly mean NH 4 + /NO 3 − ratio of 1.1) has been identified.Regarding airmasses, high VWM concentrations of nitrogeneous compounds are found in Saharan, Continental and Local categories.In link with synoptic situation and meteorological conditions, the highest VWM values were registered in the hybrid anticyclonic weather types: AN, ANE, AE and ANW for nitrate (with values of between 40 and 94 μeqN/L) and AN and ANE for ammonium (with values in excess of 50 μeqN/L).A gradual increase of VWM of NH 4 + over the years of study has been detected and seems to be mainly linked with the intensification of the pig farming and associated emissions in the Andalucia region and with the increase in the number of Sahara 500 m events.
At the EMEP station in Víznar, no acid rain problems were detected, and the calcium, sodium, ammonium and magnesium ions seem to be dominant in the neutralisation of the rainwater.Nitrate and sulphate (r = 0.83), from antropogenic sources, seem to have a similar chemical behaviour in precipitation, as well as to the co-emission of their precursors SO 2 and NO x .The highest VWM for sulphate was registered in the hybrid anticyclonic weather categories, with values of around 63 μeqS/L for AN, ANE and ANW.
Chlorine and sodium, with a marine origin, register the highest volumetric concentrations if the wind component is from the west both with an anticyclone and a cyclone.The three classes of marine origin have the lowest concentrations of the nine categories that were analysed, with the Mediterranean class being the most contaminated.The larger number of arrivals of masses of maritime air recorded during winter explains the maximums VWM of chlorine and sodium during this season.
Ca 2+ and Mg 2+ with a main crustal origin, present higher concentrations in the presence of hybrid anticyclonic weather types (between 76 and 183 μeq/L for Calcium and between 21 and 58 μeq/L for magnesium).High cation concentrations (mainly Ca 2+ , with mean volumetric concentrations of up to 150 μeq/L) characterises the chemical composition of the rainwater associated with incursions of Saharan dust (at the 3 levels) as well as high pH levels.The Saharan 3000 m category does not appear to have a significant influence (according to the Mann-Whitney test) on the chemical composition of the precipitation.
The chemistry of the rainwater is dominated by Ca 2+ and SO 4 2− , with VWM of 55 μeq/L and 34 μeqS/L during the studied period (2002)(2003)(2004)(2005)(2006).During summer, the majority of the maximum average seasonal concentrations was recorded (except for Cl − and Na + ) due to the minimum dilution experimented; a large number of the precipitation events lefting behind a small precipitated volume.It is important to emphasize the high VWM recorder for calcium during summer (206 μeq/L) due to the frequent intrusions of Sahara dust.
This study points out the relevance of long range transport on atmospheric pollution and contributes to the knowledge on rainwater quality in background stations in Europe.Considering both meteorological and airmasses studies seems to be a very useful tool for characterizing and understanding precipitation chemical composition.Further studies are needed in order to complete this data with aspects such as organic deposition (mainly N-deposition) or dry deposition in order to create a complete database that permits to evaluate modelling exercises and improve knowledges about future environmental and human health impacts.

Fig. 4 .
Fig. 4. a) Precipitation records (mm) and number of cases, b) Volumetric mean (VWM) for the components analysed (in µeq/L) and c) pH and conductivity for the different air masses categories established.

Fig. 5 .
Fig. 5. a) Precipitation records (mm) and number of cases, b) Volumetric mean (VWM) for the components analysed (in µeq/L) and c) pH and conductivity for the different categories of Circulation Weather Types (CWT) established.

Fig. 6
Fig. 6(a) shows monthly VWM of inorganic nitrogen (NH 4 + and NO 3 − ) and NH 4 + /NO 3 − during the five years of study.It can be observed important variations of VWM with time, with an opposite trend to precipitation (Fig. 6(b)), indicating a dilution effect of rainwater on inorganic nitrogen concentration.This fact pointed out that quantity

Fig. 7 .
Fig. 7. Relationship between the sum of the concentrations of Ca 2+ , Mg 2+ , NH 4 + -N and Na + and the sum of SO 4 2− -S and NO 3 − -N.
and K + have the lowest values, with annual means of between 67 and 105 mg/m 2 /year and between 42 and 90 mg/m 2 /year, respectively.A similar trend is seen for the majority of the ions: a progressive decrease in the wet deposition from 2002 to 2005, then increasing in 2006, reaching the maximum values in the five years of the study for ammonium and calcium (with nearly 193 and 670 mg/m 2 /year, respectively).If we calculate the mean annual wet deposition as the arithmetic mean of the wet deposition of the five years of the study, we obtain values of 270 mgS/m 2 /year (from the SO 4 2− -S), 155 mgN/m 2 /year (from the NO 3 − -N), 165 mgN/m 2 /year (from the NH 4 + -N), 212 mgNa + /m 2 /year, 89 mgMg 2+ /m 2 /year, 542 mgCa 2+ /m 2 /year, 422 mgCl − /m 2 /year, 0,25 mgH + /m 2 /year, and 69 mgK + /m 2 /year.Focusing on nitrogen deposition, from 2002 to 2006, yearly N deposition ranged from 111 to 193 mgN/m 2 /year for NH 4 + -N (with the average of 165 mgN/m 2 /year), from 95 to 197 mgN/m 2 /year for NO 3 −

Table 1 .
Volumetric mean (VWM) and standard deviation (VWSD) recorded during each of the years of the study and during the whole period(2002)(2003)(2004)(2005)(2006).All of the concentrations included are in μeq/L with the exception of the pH and conductivity (μS/cm).

Table 2 .
Mean seasonal precipitation (P in mm), number of precipitation events (N) and seasonal volumetric concentrations (VWM) in µeq/L (except for the pH and conductivity -in µS/cm) with their standard deviations (VWSD) for each of the parameters studied.

Table 3 .
Volumetric weighted mean (VWM) registered in different locations around the worl (µeq/L except pH).

Table 5 .
Principal component analysis of the different parameters analysed in the precipitation collected at the EMEP station in Viznar during the period 2002-2006, as well as the different parameters used to calculate the weather types.

Table 6 .
Enrichment factors of the different ions analysed in the rainwater with respect to the crustal material and seawater.
26 μeq/L and 4 μeq/L.The lowest precipitated volume was recorded in 2005 (276 mm) and the highest concentrations of Mg 2+ and Ca 2+ , as well as the highest conductivity, with values of 20 μeq/L, 77 μeq/L and 18 μS/cm, respectively.In 2006 the highest average concentration of NH 4 + was recorded, a high concentration of Ca 2+ and the highest pH, while in 2002 the highest concentration of H + and Na + was recorded.