Trace Element Distribution in Surface Winter Snow of the Eastern Alps (Italy) in Relation to Meteorological Conditions

Carlo Barbante^*, Paolo Cescon, Gabriele Capodaglio (Centro di Studio sulla Chimica e le Tecnologie per l’Ambiente - CNR and Dipartimento di Scienze Ambientali, Università di Venezia, Italia), Giulio Cozzi, Paolo Gabrielli (Dipartmento di Scienze Ambientali, Università di Venezia, Italia), Clara Turetta (Centro di Studio sulla Chimica e le Tecnologie per l’Ambiente - CNR Venezia, Italia), Sandro Torcini (Dipartimento Ambiente, ENEA, Roma)

ABSTRACT

A research program was carried out in the Eastern Alps (Italy) in order to study the chemical composition of the low troposphere using the snow as an indicator of the tropospheric aerosol content. Samples were collected at 21 Alpine sites, close to manual and automatic meteorological stations, between December 1997 and April 1998. Several trace elements (Ag, Ba, Bi, Cd, Co, Cr, Cu, Fe, Mo, Mn, Pb, Sb, Ti, U, Pt, Pd, Rh, V and Zn) and some major ions (S0₄^2‑, N0₃^-, Ca²⁺, Mg²⁺, K⁺, Na⁺ and Cl^-) were analysed by inductively coupled plasma sector field mass spectrometry (ICP‑SFMS) and by ion chromatography, respectively.

The results show that the aerosol entrapped in the snow is characterised by both an anthropogenic influence, originating from the Alpine valleys and nearby industrialised regional areas and a natural component consisting of marine and crustal influences. The hypothesis advanced to explain these results is the occurrence of atmospheric transport on a local and regional scale, due not only to convective vertical flows but also to the lifting produced by orographic winds and by the turbulence generated by synoptic wind.

INTRODUCTION

Study of the air quality in Alpine regions by means of chemical analysis of fresh snow, collected at high altitude, represents a useful and original approach to evaluation of the environmental contamination of these areas. This is important in the case of trace elements, in order to understand the atmospheric wet and dry deposition pathway of these micropollutants. The interest of researchers in these chemical species is due to their toxic effect on human beings, animals and other organisms, and to define their relationships in the biosphere. This kind of study is not so easy to carry out because of the lack of suitable methodologies for collecting and analyzing this kind of matrix in which trace elements are at or below the pg g^-1 level. These very low concentrations need special field sampling procedures to prevent sample contamination. The sampling sites were located in the Province of Trento (12) and in the Veneto Region (9): the sites were mainly located a long way from rural areas and for the most part close to ski areas at an average altitude of 1830 m. The remoteness of the sampling points is the most important criterion in order to avoid contamination from local sources of gas and particle emissions (roads, villages, tourist stations, Alpine huts and even artificial snow producing devices)

Methods

In the recent years many data regarding heavy metal in precipitation have considered unreliable because careful precautions were not taken to avoid sample contamination. Sample collection was therefore conducted according to the stringent contamination free procedures used to sample the surface snow in Greenland and Antarctica (Boutron, 1990). During the 1997-1998 winter season, 366 samples of superficial snow were collected at 21 sites in the eastern Alps.

The surface sampling was carried out by plunging ultra-clean wide-mouth bottles (500 ml) of LDPE (low density polyethylene) directly into the snow, downwind of the operators. The sampling containers were then capped, sealed in double PE bags and stored frozen in a cold room (-20°C) until analysis in laboratory. During the sampling procedures operators wore special clean-room clothing and polyethylene gloves in order to avoid contamination from themselves (Barbante, 1999). LDPE sampling bottles were previously acid cleaned in a clean room under a class 100 laminar flow hood (Barbante, 1997), using ultrapure water and ultrapure concentrated HNO₃.

The samples were treated and melted at room temperature in the LDPE sampling bottles in the clean laboratory inside a class 100 laminar flow clean bench. A 5 ml aliquot was taken in a 10 ml ultraclean LDPE test-tube for the determination of very low concentrations (from ppt to ppb level) of 16 metals ( Ti, V, Cr, Mn, Fe, Co, Cu, Zn, Mo, Ag, Cd, Sb, Ba, Pb, Bi and U) with a Finnigan MAT Element (Finnigan MAT, Bremen, Germany) Sector Field Inductively Coupled Plasma Mass Spectrometry (Barbante, 1999). Every aliquot was acidified (2%) with ultrapure concentrated HNO3.

Another 30 ml aliquot was taken in a 100 ml ultraclean LDPE bottle for the determination of much higher concentrations of the major ions (from ppb to ppm level) SO₄^2- , NO₃^-, Ca²⁺, Mg²⁺, K⁺, Na⁺, Cl^- by Ionic Chromatography (IC) using a DIONEX AS5 column for the anion determination and a DIONEX CS12 column for the cation determination.

RESULTS AND DISCUSSION

Considering the 23 elements and ions analysed, more than 8000 potential data of concentration in the snow were obtained. For this reason attention was focused mainly on the overall structure of the data set trying to extract the most essential hidden information. During the period mentioned, snow events were mainly observed in December 1998, in the first half of January 1998, at the end of February and in April 1998. February and March 1998 were characterised by stable weather and an almost total absence of precipitation. Because of these meteorological conditions, wet deposition of aerosol occurred in the first, in the central and in the last part of the sampling period whereas dry deposition took place in the period from 21 January to 18 February and from the beginning to 20 March. As dry deposition was a more continuous aerosol deposition process, higher concentrations were in general observed during February and March and low concentrations in the periods characterised from frequent snow precipitation (Fig. 1). The spatial variability of the mean concentrations of the chemical species proved very high, reflecting the heterogeneity of the sampling sites due to different altitude, geographic position and isolation from anthropic aerosol sources. In general lower concentration values were revealed in the sampling sites located above 2000 m., whereas higher values occurred in the samples from the station in the foothills of the Dolomites in the Veneto.

Figure 1 Temporal profile for lead (December 1997-April 1998) at Malga Bissina (TN)

The whole data set of concentration measurements was subjected to a factor analysis in order to highlight the relations between elements and ions. Of the various factor analysis methods, principal component analysis (PCA) was adopted. This method applied to the entire chemical data set brings out three factors: the first comprises all the trace metals, the second Cl^-, Na⁺, SO₄^2-, NO₃^- and the third Mg²⁺ and Ca²⁺. It is relevant to underline that some metals such as Ti, U, Fe, Mn obtained rather high scores even on the third factor and in fact applying a cluster analysis to the factor score matrix, the metals split to two sub-groups respectively composed of Ti, U, Co, Cr, Cu, Mn, Fe, Ba and Cd, Ag, Zn, V, Sb, Bi, Pb, Mo (Fig. 2).

Figure 2 Cluster Analysis which shows the correlation among the chemical species (See text for explanation).

Moreover the cluster analysis also splits the second factor into two distinct groups: SO₄^2-, NO₃^- and Na⁺ and Cl^-. An explanation of the first factor splitting can be obtained by comparing the metal concentrations in the snow with the average concentration measurements in carbonates, the most characteristic background rocks of the eastern Alps. The average percentage of crustal contribution was as follows: Mn 100%, U 91%, Ti 65%, Fe 50%, Cr 32%. The average crustal contribution for Ag, Cd, Co, Sb, Mo, Cu, Pb, Ba and Zn was negligible (< 3%). This cross evidence clarifies rather well the reason for the connection between the third factor (Ca²⁺, Mg²⁺) and some distinct metals that seem to assume a “crustal character” and in particular Mn, U, Ti and Fe which achieve the highest carbonate contribution and at same time quite high scores in the third factor. The factor analysis, concurrently to the cluster analysis, groups some metals which, thanks also to the carbonate contribution calculation, seem not to be referable to a carbonate origin. Of these metals, Pb and Cd in particular can be referred to an “anthropogenic origin”, firstly because the human emission contribution is greatly prevalent and secondly because the provenance was already widely recognised as anthropogenic in other similar studies carried out in the Alps (Barbante, 1999)(Van de Velde, 1998). The cluster analysis also split the second factor combining Na⁺, Cl^- and SO₄^2-, NO₃^-. The good relation between Na⁺ and Cl^-, concurrently with the comparison with the marine concentration ratio (Cl^-/Na⁺ = 1.803), shows a marine origin probably referable to the nearby Adriatic basin. The association between SO₄^2- and NO₃^- suggests an anthropogenic origin: in fact the emission of the respective original compounds is much more predominant than the natural sources (Corinnair 1990). The non correlation among the chemical variables and in particular the horizontal transport factors on an extended spatial scale (wind velocity in the free atmosphere) suggests the hypothesis of a local and/or regional origin of the aerosol deposited elements and ions. It appears in fact that the very industrialised Po Valley may affect the more exposed Alpine slope concurrently with the action of other very diffuse sources of emission (roads, tourist station, urban centres) within Alpine valley. The mechanism of transport of the aerosol particles from the low altitude to the superficial snow mantle of the high Alpine slopes could be due, in addition to the weak convective turbulence inside the boundary layer, to the orographic winds and to the wind turbulence generated during the sampling period by synoptic wind action.

REFERENCES

Boutron CF (1990), Fresenius J. Anal. Chem. 337: 482-491.

Barbante C, Bellomi T, Mezzadri G, Cescon P, Scarponi G, Morel C, Jay S, Van de Velde K, Ferrari C and Boutron C F (1997), J. Anal. At. Spectrom. 12: 925-931.

Barbante C, Cozzi G, Capodaglio G, Van de Velde K, Ferrari C, Boutron C F and Cescon P (1999), J. Anal. At. Spectrom. 14: 1433-1439.

Corinnair 1990, Inventory of the European emission

Van de Velde K, Boutron C F, Ferrari C, Bellomi T, Barbante C, Rudnev S and Bolshov M (1998), Earth Planet. Sci. Lett. 164: 521-533.