Introduction

Rainwater composition depends on two physical phenomena, the rain-out and wash-out (Berner and Berner 1996). Rain-out is the formation of raindrops from condensation nuclei; wash-out is the removal by formed raindrops. So chemical elements and compounds present in atmosphere, like gases or solid particulate matters, can be transported by the wind as aerosols and travel also long distances before being deposited by rain, influencing chemical parameters of soil, superficial water and vegetation (Niu et al. 2014; Honório et al. 2010). Rainwater chemical composition is due to natural and anthropogenic mineralization processes (Berner and Berner 1996) and so it can provide valuable information on sources of local air pollution too (Montoya-Mayor et al. 2013; Vázquez et al. 2003). Thus, the chemical investigation of rainwater is useful to trace different sources of some local atmospheric pollutants (Oliveira et al. 2012). Chemical composition of gases together with solid particulate matters in lowest portion of Earth's atmosphere depends on the emission sites, on the topographic height and on the meteorological conditions (Santos et al. 2011; Flues et al. 2002).

In urban areas rainwater composition changes from site to site due to the effect of local sources (Migliavacca et al. 2005; Kulshrestha et al. 2003; Flues et al. 2002). The chemical composition of rainwater harvested at soil level can notice relevant air pollution dynamics. For instance, the combustions of vehicular traffic and industries could be the dominant anthropogenic local sources of trace elements in rainwater (Cheng et al. 2011; Hu and Balasubramanian 2003; Kaya and Tuncel 1997). Trace elements from precipitation accumulate in the biosphere and may cause adverse human health and environmental effects (Báezet al. 2007; Barrie et al. 1987), thus the studies of trace elements in rainwater have increased in many countries (Báezet al. 2007; Al-Momani 2003; Pike and Moran 2001).

In the most of the cases, the same chemical element or compost is enriched by more than one mineralization process. Therefore, two or more mineralization dynamics could overlap and as a consequence, the concentration of an element or compost depending on both or more processes could not be traced.

In the proposed investigation single rainfall events have been collected separately in the southern Plain of the region known as Campania (located between Caserta and Napoli provinces, Southern Italy). This area is highly urbanized and industrialized but there is a high activity of intensive agriculture too. It is characterized by the presence of heavy traffic and industries but recently it is known as “Land of Fires” due to illegal uncontrolled burnings (National Institute of Health), so much so that by now it is even studying with a political ecology of health approach too (D’Alisa et al. 2017; Iengo and Armiero 2017).

The aim of this study is detect and separate the rainy events in which are predominant specific mineralization processes, by this to be a first step to show a relationship between their rainwater compositions and the local air pollution, if any. Overall, there is evidence of connections between rainwater composition and its sources, including natural processes as well. In fact, in the future eventually there should be a study on trace elements in air, to compare with the values obtained in rainwater.

Materials and methods

Study area

All the sampling sites are long the same perpendicular to the coastline. They are situated in the north hinterland of Naples and in the city of Caserta, a town of almost 80,000 people in the north of Naples with a province of less than one million people.

This area presents a typical Mediterranean climate and its dominant winds mainly come from Balkans (from east) or from the sea (from west). The first case is due to the atmospheric low-pressure centre at about 300 km far south-southeast and the second one at about 500 km far northwest (European satellite sat24), so that air masses move around low-pressure centers with a counterclockwise sense of course (Fig. 1).

Fig. 1
figure 1

Low pressure centers (L) around that air masses move on large scale

Full size image

The province of Naples instead have more than 3 million people and here there are two sampling sites relatively far from relevant anthropogenic emissions, one near the sea (Lago Patria—LP) and another one in the middle between Lago Patria and Caserta (Sant’Antimo—SA). Naples is on the west coast of the south of Italy and Caserta is in the north of Naples about 30 km from the sea and about 20 km from Naples. In the city of Caserta there are other two sampling sites relatively far from relevant anthropogenic emissions, one in the city-centre (San Benedetto—SB) and another one on Tifata Mt. (Casertavecchia—CV). Figure 2 shows the sampling sites.

Fig. 2
figure 2

Study area map and sampling sites layout

Full size image

Local winds are mainly influenced by sea breeze with its proximity and so their direction is perpendicular to the coastline, like the sampling sites. The soil origin is mostly volcanic due to the proximity to Mount Vesuvius volcano and so the substrates are both carbonated and volcanic. The Campanian Plain covers all the land from the sea to Caserta, that is at an altitude of about 68 m a.s.l. and the first hill is Caserta Vecchia (on Tifata Mt.), that is at an altitude of about 400 m a.s.l.

The area between Naples, Caserta and Sant’Antimo is highly urbanized, the area between Sant’Antimo and Lago Patria has a long tradition of agriculture and Sant’Antimo is in the middle of the famous worldwide so-called “Land of fires” due to illegal uncontrolled burnings. For example, the pictures of Fig. 3 were taken from Casertavecchia seeing Naples in different days of 2017 summertime. The second picture in Fig. 3 also shows very well how the local winds can push the lowest air masses and so the smoke plume from sea to inland during sunny hours.

Fig. 3
figure 3

Pictures taken from Casertavecchia seeing Naples in different days of 2017 summertime

Full size image

Methodology

The minimum intensity collectable was related to rainy events of about 1 mm/m2. Each single rainy event was collected using polyethylene funnels set up shortly before and get off soon after each rainy event.

Rainwater samples were divided in the field into two aliquots. The first was filtered, with a 0.45-μm Millipore cellulose acetate membrane filter, for the analysis of anions and cations. The second was also filtered (with a 0.45-μm Millipore cellulose acetate membrane filter) and acidified (1% of Suprapur HNO3) for the analysis of trace elements.

Electrical conductivity (EC) and pH were determined just after the rainy event by a portable instrument (HANNA—HI 991300). Total alkalinity (as HCO3−) was determined in the field by titration with 0.01 M HCl and methyl orange as indicator.

The analytical equipment IC together with ICP-MS allows to detect almost the totality of inorganic solutes in a concentration range compatible with usual rainwater compositions. Cations (Na+, K+, Mg2+ and Ca2+) and anions (Cl−, NO3− and SO42−) were analyzed by ion chromatography and trace elements (B, Al, Si, Mn, Fe, Cu, Zn, Rb, Sr and Ba) by ICP-MS both at the Department of Environmental, Biological and Pharmaceutical Sciences and Technologies (University of Campania “Luigi VANVITELLI”) using procedures of Tommasone et al. (2011), Cuoco et al. (2013). The detection limits were 0.05 mg/l for all the cations and anions by ion chromatography and 1.0 µg/l for all the trace elements by ICP-MS. Since the results are expressed in mol/l, the detection limits indicated in the Table of Results have been before converted from mg/l to mol/l for each parameter.

Analytical precision and accuracy for repeated analyses of the rainwater and of the international and internal standards were better than 5%.

Statistical treatment of data

The statistical treatment of the samples was carried out with Statgraphics version XVII (Statgraphics Graphics Corporation, ST.SC., USA).

After calculating mean, median, maximum and minimum values, the data were transformed to achieve normality. In addition, data were processed by means of Factor Analysis (FA), applying the Varimax rotational technique to facilitate easier interpretation. Varimax rotation was applied because orthogonal rotation minimizes the number of variables with a high loading on each component and facilitates the interpretation of results. The purpose of the analysis is to obtain a small number of factors which account for most of the variability in the variables set. In some cases, samples have been grouped using dendrograms.

Results and discussion

Each rainfall event of 1 mm/m2 minimum intensity were collected from May 2016 to January 2017 in all the sampling sites. 45 samples were collected in Casertavecchia and in San Benedetto, 42 samples in Sant’Antimo and 36 samples in Lago Patria.

Meteorological conditions and mineralization of rainy events along the sampling period seasons

In all the sampling sites, the frequency of the rainfall was similar throughout the seasons, except for the months of August and December, when they were almost absent. Their mean intensity over the whole sampling period was about 10 mm/m2 and it was lower during the hot seasons (spring–summer time) than in the cold ones (data from Regional Agro-meteorological Center, Campania Region). Instead, the mean EC was higher during the hot seasons than in the cold ones in all the sampling sites, as well as decreasing with increasing rainy intensity.

First data processing

All the analytical concentrations, that are expressed in mg/l for the major elements and in µg/l for the trace elements, have been transformed in mol/l.

Table 1 shows the mean, the median, the maximum and the minimum values of all the measured parameters in each sampling site. Acidity (evaluated from pH measurements), EC and major elements often showed increasing trends from Casertavecchia—San Benedetto—Sant’Antimo up to Lago Patria, strongly visible for Cl− and Na+; it is worth to note that NO3− and SO42− high concentrations related to limited number of rainy events were detected. The NO3− and SO42− mean and median values follow the above-mentioned trends but the maximum values attain concentrations really anomalous for the most of worldwide urban rainwaters (Berner and Berner 1996). Trace elements showed comparable median concentrations except for Cu in Sant’Antimo where it was over three times much higher than in all the other sampling sites.

Table 1 Distributions values of all the measured parameters (BDL = below detection limit)
Full size table

Therefore, all the singular element or compound distributions in each sampling site have been as normalized as possible and out-performed with Bias and Kurtosis values between − 2 and + 2.

The factorial analysis on the major elements for each sampling site has been conducted and everywhere it has obtained Kaiser–Meyer–Olkin values > 0.6. It showed that two similar factors are in all the sampling sites, except in Lago Patria, where there is only one (see Fig. 4). In fact, the first factor is mainly related to HCO3− and Ca2+ while the second one is mainly related to Cl− and Na+ in all the three sampling sites. Moreover, NO3−, SO42− and K+ are only related more to the first factor than to the second one while Mg2+ only more to the second than to the first in all the three sampling sites. The first factor represents the continental character and the second one the marine for the nature of their main substances. In Table 2, weight values by major elements factorial analysis can be seen.

Fig. 4
figure 4

Graphics of major elements factorial analysis

Full size image
Table 2 Weight values by major elements factorial analysis
Full size table

Sometimes the signs of the same factor weights are opposite between two sampling sites due to arithmetical reasons for the different functions applied for the normalization.

The factorial analysis suggested to divide database in three groups of data, depending on the arithmetical difference among the absolute values of their two weights for each sample: the marine, the intermediate and the continental group. The intermediate group consisted in the samples with an arithmetical difference between 0 and 1 among the absolute values of their two weights. The continental and marine characters of the rainy events did not depend on seasonality in any of the three sampling sites.

Relations among NO3 −, SO4 2− and all the trace elements with both NO3 − and SO4 2− for all the groups of data

The high concentrations of NO3− and SO42− in all the sampling sites, also simultaneously, betray the presence of anthropogenic processes. Recently it showed that their origins are anthropogenic also at lower concentrations than these ones. In particular, the maximum values of NO3− and SO42− in the study of Mimura et al. (2016) are less than 6 mg/l and their origin is vehicular traffic. In the study of Chon et al. (2015) the origins of NO3− and SO42− are fossil fuels; the mean and the maximum values of NO3− are, respectively, 2.3 mg/l and 12.4 mg/l while the mean and the maximum values of SO42− are, respectively, 4.4 mg/l and 19.5 mg/l. SO42− mean value in the study of López et al. (2015) is less than 2 mg/l and its origins are biomass combustion and volcanic emissions. NO3− maximum weekly value in the study of Moreda-Piñeiro et al. (2014) is less than 7 mg/l and its origins are fertilizers, coal combustion and vehicular traffic. The maximum values of NO3− and SO42− in the study of Al-Momani (2003) are less than 12 mg/l and their origins are vehicular traffic and combustions.

Thus, R coefficients among NO3−, SO42− and all the trace elements with both NO3− and SO42− have been calculated for all the groups of data in each sampling site, indicating the number of the events and the relative p value too, where a p value < 0,05 claims the statistical significance. Moreover NO3− and SO42− sample values suggested a bimodal distribution: using the NO3−/(2*SO42−) ratio imbalance each group of data has been separated in other two groups of data that have generally developed higher R coefficients among NO3− and SO42−, but above all among trace elements with both NO3− and SO42− in the continental groups of data. This ratio is usually lower in regions impacted by sources such as oil refineries, petrochemical industries and coal-fired power plants (Migliavacca et al. 2005; Flues et al. 2002).

Sometimes they have not been calculated because there are few events and so the relative statistics was impossible. Moreover, also in this case sometimes the signs of the same R coefficient are opposite among two sampling sites due to arithmetical reasons for the different functions applied for the normalization. In all the sampling sites, 2*SO42− > NO3− rainy events were much more numerous during the cold seasons than during the hot ones.

All the calculated R coefficients among NO3−, SO42− and all the trace elements with both NO3− and SO42− for all the groups of data in each sampling site are in the Electronic Supplemental Material (ESM), but following the tables with the main results are shown.

In Caserta Vecchia, the ratio among the number of events with 2*SO42− > NO3− and with NO3− > 2*SO42− is pair to about 36%. R coefficients among NO3− and SO42− show increasing trend from whole database to marine, to intermediate and to continental so like from NO3− > 2*SO42− to 2*SO42− > NO3− groups of data. Generally, R coefficients among the trace elements with NO3− and SO42− follow this trend as well and the only groups of data with high R coefficients among the trace elements with NO3− and SO42− are those with 2*SO42− > NO3− both of marine and continental; these high R coefficients are few.

In San Benedetto, the ratio among the number of events with 2*SO42− > NO3− and with NO3− > 2*SO42−stands at 50%. R coefficients among NO3− and SO42− show increasing trend from whole database to both intermediate and continental, so like from NO3− > 2*SO42− to 2*SO42− > NO3− groups of data. The main result is that the group of data with 2*SO42− > NO3− of continental shows too many high R coefficients among all the trace elements with NO3− and SO42−, in fact these are with all the trace elements except with Rb (see Table 3).

Table 3 R coefficients among NO3−, SO42− and all the trace elements with both NO3− and SO42− for the group of data with 2*SO42− > NO3− of continental in San Benedetto
Full size table

In Sant’Antimo, the ratio among the number of events with 2*SO42− > NO3− and with NO3− > 2*SO42− is stands at 68%. R coefficients among NO3− and SO42− show increasing trend from whole database to marine and to continental; R coefficients among NO3− and SO42− show increasing trend for whole database and continental from 2*SO42− > NO3− to NO3− > 2*SO42− groups of data. The main result is that the group of data with NO3− > 2*SO42− of continental shows many high R coefficients among all the trace elements with NO3− and SO42− (see Table 4), but also the groups of data with NO3− > 2*SO42− of marine and with 2*SO42− > NO3− both of intermediate and continental show several ones.

Table 4 R coefficients among NO3−, SO42− and all the trace elements with both NO3− and SO42− for the group of data with NO3− > 2*SO42− of continental in Sant’Antimo
Full size table

In Lago Patria, the ratio among the number of events with 2*SO42− > NO3− and with NO3− > 2*SO42− is pair to about 80%. R coefficients among NO3− and SO42− show increasing trend from whole database to NO3− > 2*SO42− and to 2*SO42− > NO3− groups of data. Both of the groups of data with NO3− > 2*SO42− and with 2*SO42− > NO3− show several high R coefficients among all the trace elements with NO3− and SO42−.

Detailed statistics for particular groups of data

The trace elements results mainly showed only that Cu concentrations in Sant’Antimo much higher than in all the other three sampling sites, R coefficients among NO3−, SO42− and all the trace elements with both NO3− and SO42− for the group of data of its five highest Cu concentrations events have been calculated (see Table 5). These events can be divided in two groups, one at the beginning of May and another one in the second half of July. The events of each new group are all consecutive and all with NO3− > 2*SO42−. Moreover, the two groups began after periods of drought about 1 and 3 weeks, respectively.

Table 5 R coefficients among NO3−, SO42− and all the trace elements with both NO3− and SO42− for the five highest Cu concentrations events in Sant’Antimo
Full size table

The main result is that there are too many high R coefficients among all the trace elements with NO3−and SO42−, in fact these are with all the trace elements except just with Cu and only because it has not linear correlations. The dendrogram of NO3−, SO42− and all the trace elements (see Fig. 5) and the dendrogram of Cu alone (see Fig. 6) produce the different results like as follows.

Fig. 5
figure 5

Dendrogram for the concentration values of NO3−, SO42− and all the trace elements obtained during the five events in Sant’Antimo with the highest Cu concentrations

Full size image
Fig. 6
figure 6

Dendrogram for the concentration values of Cu obtained during the five events in Sant’Antimo with the highest Cu concentrations

Full size image

In particular, the first dendrogram confirms the groups division in two different periods, with gradual relevance along each singular period. In fact, the group of the beginning of May is composed by samples with ID pair to 4, 5 and 6; the group of the second middle of July is composed by samples with ID pair to 16 and 17.

The second dendrogram shows the different correlation of Cu compared with the dendrogram before explained and in particular that the first samples of each period have the same relevance, as like the second ones have another different same relevance.

Since in spring–summer time the only period that have the minimum number of events with 2*SO42− > NO3− in San Benedetto is during the months of June and July (only 1 event on 11 events, in particular on the 18th of June), for first the factor analysis has been conducted (see Table 6).

Table 6 Data of major elements factorial analysis in San Benedetto during June and July
Full size table

The main result confirms that NO3− and SO42− are more related to continental factor although in this case the first factor is the marine one as a consequence, R coefficients among NO3−, SO42− and all the trace elements have been calculated also for its continental group of data with NO3− > 2*SO42− (see Table 7).

Table 7 R coefficients among NO3−, SO42− and all the trace elements with both NO3− and SO42− for the continental events in San Benedetto during June and July with NO3− > 2*SO42−
Full size table

It shows a high R coefficient among NO3− and SO42− but above all it does not show high R coefficients among all the trace elements with NO3− and SO42−, except only for Cu and Sr although with slightly high p values.

Likewise R coefficients among NO3−, SO42− and all the trace elements have been calculated in Sant’Antimo for its continental group of data with 2*SO42− > NO3− and with Cu concentrations < 5 µg/l (see Table 8). These six events never happened in spring–summer time except one, in particular on the 12th of May.

Table 8 R coefficients among NO3−, SO42− and all the trace elements with both NO3− and SO42− for the continental events in Sant’Antimo with Cu concentrations < 5 ppb and with 2*SO42− > NO3−
Full size table

It confirms the high R coefficient among NO3− and SO42− but above all it shows a strong increasing trend of R coefficients among NO3−, SO42− and all the trace elements with NO3− and SO42− if compared with the whole continental group of data with 2*SO42− > NO3−.

So, since Rb is the only trace element without high R coefficient with NO3− and SO42− in San Benedetto during the continental events with 2*SO42− > NO3−, R coefficients among all the trace elements have been calculated in Sant’Antimo for its continental group of data with NO3− > 2*SO42− (see Table 9).

Table 9 R coefficients among all the trace elements for the continental events in Sant’Antimo with NO3− > 2*SO42−
Full size table

It confirms that Rb, and also B, are the trace elements with the most numerous high R coefficients with all the other trace elements and also in this case with all the trace elements except with Cu; the highest R coefficient is right among Rb and B.

Moreover, Rb and B show their highest concentrations in the spring–summer time, just like Cu, and often same trends too (see Fig. 7).

Fig. 7
figure 7

B, Cu and 10*Rb concentration (10–5 mol/l) trends in Sant’Antimo

Full size image

Main mineralization process

The sampling sites have very specific characteristics, each one differs from all the others one, despite their closeness. Therefore, the process that causes the most of these rainwaters chemical composition is the vertical wash-out of the atmospheric layers closest to the soil, like in the most of worldwide urban rainwaters. This is confirmed also by, increasing trends of the acidity, EC and major elements from Casertavecchia—San Benedetto—Sant’Antimo up to Lago Patria, strongly visible for Cl− and Na+ certainly for the gradual proximity to the sea. Moreover, in all the sampling sites other confirms one from the EC decreasing with increasing rainy intensity and from some rainy events that showed all the concentrations at least 10 times lower than their mean concentrations or even below detection limit.

Probable polluting processes

The great statistics, the bibliographic references and the morphology and activities of the investigated region have suggested the dynamics above, so individuating the following sources.

Since fuels utilized by vehicular traffic and by Industry are now low in Sulfur and since the increasing trend of the ratio among the number of events with 2*SO42− > NO3− and with NO3− > 2*SO42− from Caserta Vecchia (36%)—San Benedetto (50%)—Sant’Antimo (68%) up to Lago Patria (80%), probably the processes that cause 2*SO42− > NO3−conditions are the illegal uncontrolled burnings, in effect more and more covering the area that stretches from San Benedetto to Sant’Antimo and to Lago Patria. These processes develop NO3− and SO42− R correlations with all the detected trace elements, except with Rb.

In fact, Iovino et al. (2009) supposed that BTEX are downwind of sea breeze and so they come from seaside during sunny hours and stop in Caserta centre zone because blocked by Tifata Mt., that is the first hill from the sea. Moreover, in the centre of Caserta there is a 18th-Century Royal Palace, that is in the World Heritage List of UNESCO since 1997, and Iovino et al. (2014) also have shown that its inside atmosphere is effected by external sources of some volatile organic pollutants too, that with time could entail a significant risk also for works of art.

Similarly, the processes that cause NO3− > 2*SO42−conditions could be vehicular traffic, Industry and above all from Sant’Antimo to Lago Patria agriculture. This last process develops NO3− and SO42− R correlations with all the detected trace elements, above all with Rb and B while not with Cu only because in this case they are not linear, but Cu relative concentration is always the highest because for example it is only linked to pesticides and/or to relative soil handling, eventually also for its volcanic composition. This would be confirmed also by the periods and trends of B, Cu and Rb highest concentrations in Sant’Antimo, typical of intensive agricultural activity.

Moreover, sea breeze has a lavage effect.

Conclusions

The experimental design showed to be a first step very suitable for locating and detailing the hardest environmental risks despite geographical extension and complexity of the investigated region. In fact, in the future there should be a study on trace elements in air, to compare with the values obtained in rainwater. All the trace elements showed high R correlations with both NO3− and SO42− for specific individuated groups of data, as well as due to specific probable polluting processes.

Caserta Vecchia shows a relatively good state compared with the other three sampling sites probably thanks to absence of strong anthropogenic activities and to its geographical position. San Benedetto mostly suffers from the pollution eventually caused by illegal uncontrolled burnings, highly involving the area between Sant’Antimo and Lago Patria, because it is downwind of sea breeze during sunny hours and so air masses of the smoke plumes stop because blocked by Tifata Mt., that is the first hill from the sea. Sant’Antimo shows the worst relative state among all the sampling sites eventually due to the presence both of illegal uncontrolled burnings and agriculture, that create a statistical disturbance among them, as like Lago Patria, which, however, is saved by sea proximity and, therefore, by lavage effect of sea breeze. Heavy traffic and Industry seem not to cause statistically significant effects if compared with both illegal uncontrolled burnings and agriculture.