1 Introduction

Anthropogenic activities are constantly altering the natural chemical composition of the atmosphere. Air pollutants are mainly produced from combustion of fossil fuels, smelting and roasting of ores for metal refining, endless urban constructions and waste incineration [84, 111, 126]. Atmospheric pollutants often have long-range transport pathways that have an important contribution on regional climate and pollution levels, since air masses passing through industrially developed regions are heavily loaded with anthropogenic pollutants [60, 115].

Atmospheric wet and dry depositions are the major processes that remove contaminants from the atmosphere [6]. The continuous process of dry deposition is predominant in areas of low precipitation [86]. In contrast, episodic wet deposition is the most effective process of atmospheric cleansing [41, 129]. The cleansing of the atmosphere by wet and dry depositions ultimately results in the transfer of both nutrients and contaminants from atmosphere into soil and aquatic ecosystem [32, 48]. The study of metals in rainwater has gained interest due to increasing air pollution and concern about the adverse environmental and human health effects of the deposited pollutant, particularly metals that become potentially toxic after entering the terrestrial and aquatic environments [95].

Chemical composition of rain water is affected by the local and regional characteristics of air pollutants [96]. Rainwater pH is controlled by the ionic composition such as Na+, K+, Mg2+, Ca2+, Cl−, HCO3−, SO42−, NO3− and NH4+ [66, 90, 106]. For example, presence of atmospheric gasses such as SO2 and NOx which are precursors of major acids H2SO4 and HNO3, respectively, results in acidic precipitation [7, 24, 74]. In contrast, CaCO3 in airborne dust and/or presence of ammonia released from anthropogenic or natural sources can neutralize acidity in rainwater [36, 98]. Thus, substances of anthropogenic origin (industry, agriculture and fertilizers, combustion of fossil fuels) such as SO2, NOx, ammonia, and toxic metals, and those of natural origin (sea salt and terrestrial aerosols) such as Na+, Mg2+, Ca2+, Cl− [14] are present in atmospheric precipitation.

Globally, studies pertaining to geochemistry of rainwater have gained tremendous momentum primarily due to two reasons [9, 29, 64, 73, 124]. First, due to the toxicity of certain metals (such as Pb, Cd, Cr, As, Hg), their capability to bioaccumulate and biomagnify and carcinogenic properties, high concentrations of such metals in precipitation are perilous for both terrestrial and aquatic ecosystems [62, 92, 122]. Hence understanding the chemical composition of rainwater, particularly the pollutants are important for the environment and ecosystem. Second, water shortage has become a pressing issue nowadays, and rainwater harvesting (RWH) is considered a viable water supply for urban regions [12, 38]. In many regions, RWH is the only sustainable way to obtain drinking water and used for crop irrigation and the recharging of aquifers [51, 77]. Thus, chemical characterization of rainwater is required to determine the treatment necessities.

There have been numerous studies on the major ion composition of rainwater in India [17, 39, 53, 67, 97, 109, 114, 119] to understand the undergoing processes such as pH control, “in cloud” and below-cloud scavenging and understanding natural verses anthropogenic sources ( [61, 66, 81, 89]. However, none of the studies quantified trace metal concentrations simultaneously in rainwater and aerosol and measure their wet deposition fluxes. Additionally, in India, compliance with the rainwater harvesting has become a necessary requirement for major construction projects to be fulfilled when receiving environmental clearance [31, 67]. Hence, knowledge about the chemistry of the rainwater that is harvested is of utmost importance. To fill in the knowledge gaps, this pilot study investigates trace metal concentrations in aerosol and rainwater during monsoon season in an Indian megacity, Kolkata. Previous studies investigated trace metals in atmospheric particulate matter (PM10 and PM2.5) from Kolkata [25, 57]; however, these studies did not quantify rainwater metal concentrations from the megacity. The main objectives of this study are to understand aerosol leaching, quantify trace metal in rainwater and understand the sources of trace metals and quantify the atmospheric wet deposition (AWD) fluxes.

2 Methods

2.1 Sampling

Aerosol and rainwater sampling was carried out in Kolkata [22° 29′ 57.53′' N, 88° 22′ 18.66′' E.], the capital of the state of West Bengal and the only megacity in eastern India. The major industries around the city are jute mills, paper and pulp industries, tanneries, textile mill, thermal power plants and oil refineries. The city experiences a tropical wet-and-dry climate [25]. Generally, monsoon commences in June and lasts till September. In 2019, monsoon was delayed by a month and commenced in July, lasted till October and supplied the city with most of its annual rainfall (Fig. 1). 72 h of air mass back trajectories at 500 m above ground level (a.g.l) terminating at Kolkata during the sampling period showed that majority of the air mass originated over Arabian Sea and blew over the Indian landmass before reaching the city (Fig. 1).

Fig. 1
figure 1

Backward air mass trajectories terminating at Kolkata back-traced for 72 h at 500 m above mean sea-level using the HYSPLIT model during the sampling days. Red trajectories are for days when aerosol (PM10) was collected and blue trajectories are for rainwater collection days. The inset bar diagram shows monthly distribution of rainfall over Kolkata in 2019

Full size image

Rainwater and aerosol were collected on the roof top of School of Environmental Studies building, Jadavpur University, approximately 24 m a.g.l. The building has no wet laboratories; hence, there are no ventilation exhausts on the rooftop to contribute blanks from acid fumes. The rainwater samples were collected in 500 ml high density polyethylene (HDPE) bottles through HDPE funnel covered with nylon net. The bottles, funnels and the net were pre cleaned with 2% reagent grade HNO3 and thoroughly washed several times with ultrapure water followed by overnight drying at 60 °C in a convection oven. The bottles were placed on the rooftop several minutes after commencement of rain and immediately removed either after rain stopped or the bottle filled up, which ever was earlier to avoid any dust or dry deposition. A total of thirteen rain water samples and two blanks were collected between July and October 2019. For the procedural blanks, the bottles with the funnels covered by net were taken to the rooftop and ultrapure water was poured inside the bottles. The blanks samples were treated similar to the rainwater samples.

A total of ten PM10 samples were collected in between rain events using a deployable particulate sampler (DPS) pump from Leland Legacy operating at 10 L/min. 24-h samples (usually starting at 8 AM) were collected on prewashed 47-mm PTFE (polytetrafluoroethylene) filters. The PTFE filters were prewashed in acid and blanks collected as described in [26] to minimize filter blanks. The filters were weighed and sealed immediately in a petri dish and stored inside a desiccator after the collection. Concentrations of PM10 were assessed gravimetrically.

2.2 Chemical analysis

Trace metal analysis of rainwater and aerosol samples was carried out at the Earth Observatory of Singapore, Nanyang Technological University (NTU), Singapore, in a Class 100 metal-free clean chemistry laboratory. Anions were analysed at the Central Environmental Science and Engineering Laboratory at NTU.

2.2.1 Rainwater

Immediately after sampling approximately 10 ml of rainwater was filtered through prewashed 0.45 µm Acrodisc® Syringe Filters with PTFE membrane into 15-ml HDPE bottles. The filtrates were acidified with several drops of 14 N ultrapure HNO3 to prevent absorption of trace metals on the bottle wall and preserved for trace element analysis. In the remaining water, pH and electrical conductivity (referred to as conductivity herein) were immediately measured and the remaining preserved for ion chromatography analysis for anions. pH was determined using Eutech pH 700 from Thermo Scientific and cross-checked using Toshcon digital pH meter. Both the pH meters were calibrated with NIST buffer sets with pH 4.01 and pH 6.86 and the electrode cleaned with distilled water before measuring the unknown samples. The difference in reading between the two instruments was within 0.09 of each other, and the average value was taken. The conductivity of the rainwater samples was measured by using a Systronics conductivity meter. Standard KCl solutions (0.01 M, 0.001 M concentrations) were used for calibrating the instrument. The conductivity and the temperature probes were rinsed thoroughly with distilled water before calibration. The different standard KCl solutions were used for calibrating the instrument before measuring the unknown rainwater sample. Temperature of the instrument was adjusted as closely as possible to 30 ± 2 ºC.

The rainwater and blank samples were analysed for anions (Cl−, SO42−, NO3−) using ion chromatography (Dionex ICS-1000 Ion Chromatography System (ICS-1000)). Before running the samples, the ion chromatography system was calibrated using a standard solution (Dionex Seven Anion Standard from Thermo Scientific). All samples were filtered through 0.45-µm cellulose acetate syringe filters before being injected into the IC column. By comparing the data obtained from a sample to that obtained from the known standard, sample ions were quantified. The final concentrations were blank-corrected using the average procedural blank concentrations. Averages of the procedural blank were 1.3% (for SO42−), 1.9% (for NO3−) and 2% (for Cl−) of the average signal intensity of the samples.

2.2.2 Particulate matter (PM10)

For determination of trace metal concentration, filters were cut into small pieces using a ceramic scissor and placed inside pre-cleaned PTFE vials. The scissor was first washed with ultrapure water followed by methanol before cutting the next sample to prevent any contamination. Trace metals were extracted in 50% (v/v) 3:1 HNO3: HF mixture. The vials were first placed in an ultrasonic bath at 60 °C for an hour. Then the vials were placed on the hot plate at 150 °C for 12 h. After cooling down the vials, caps were opened and the filter papers were washed with ultrapure water inside the vials and discarded. The digested samples were then evaporated to dryness and re-dissolved in 5 ml of 50% HNO3 and dried again. The evaporation and re-dissolving procedure were then repeated a couple of times to ensure complete evaporation of HF. Finally, the dried metal extract was dissolved in 5 ml of 2% HNO3 and transferred to precleaned vials for ICPMS analysis. Procedural blank was measured on two filters that were taken to the field and brought back and were treated in the same way as the samples. SRM 2783 of Urban Particulate Matter standard was used to test the extraction efficiency and validate the method [25].

2.3 ICP MS measurements

Trace metals (Na, Ca, Mg, Al, Fe, Mn, V, Cr, Ni, Cu, Zn, Cd, Pb) were determined in the acidified rainwater samples and aerosol extracts using ICPMS (Inductively Coupled Plasma Mass Spectrometry, Thermo Element 2, USA). High-purity multi-element standard solution (IV-ICPMS-71A from Inorganic Ventures) diluted to appropriate concentrations depending on the signal range was used to measure the metal concentrations. One ppb of indium (In) was used as an internal standard to correct for instrument drift. Samples were measured using sample standard bracketing technique. The procedural blank intensities that ranged from 1 to 3% of the average signal intensity of the samples were subtracted from the sample signal intensities. SRM 2783 of Urban Particulate Matter samples were used for method validation. Recoveries for all metals were found to be > 90% [25]. The limit of detections (LODs) of all analysed elements was calculated based on three times of the standard deviation (SD) of blank. The value of LOD ranged from 0.0023 (Pb) to 2.2717 (Ca) ppb (n = 12) (Kayee et. al., 2020). Chemical composition of all the samples (rainwater and aerosol) is given in Table S1 in the supplementary information section. Further details of the analytical procedures are documented in [25, 26] and [59].

2.4 Chemical data analysis

The measured metal and anion data were further used to calculate the enrichment factor (EF), sea salt component, and wet deposition flux (WDF) and statistical analysis.

2.4.1 Statistical analysis of the data

Pearson's correlation coefficient (r) is used to measure the strength of the association between the two variables. The correlation coefficient is considered to be statistically significant if the p values were ≤ 0.05.

To identify the probable sources of the trace elements in aerosol, principal component analysis (PCA) with varimax rotation and Kaiser normalization is used. In PCA, a multivariate data table is represented as smaller set of variates that have similar patterns in the real world such as source, atmospheric dispersion pathways, etc. Factors with eigenvalues ≥ 1 are only considered. Eigenvalues < 1 are negligible from a variance point of view. To select the significant features for the interpretation of each component, factors loadings greater than 0.5 are considered.

All statistical analyses were performed using SPSS for Windows V. 16.0.1.

2.4.2 Crustal enrichment factor (EF)

Crustal enrichment factor (EF) is often used to identify anthropogenic sources of metals over natural background. Metal ratios are generally compared with the average upper continental crustal (UCC) composition. In this study, we use UCC composition from Rudnick and Gao [93]. The EF of an element in a PM and rainwater sample is defined as: \({\text{Enrichment Factor}} = \frac{{({\text{metal conc}}{\text{./ref}}{\text{.metal conc}}.)_{{\text{aerosol/rainwater}}} }}{{({\text{metal conc}}{\text{./ref}}{\text{.metal conc}}.)_{{{\text{crust}}}} }}\).

Elements that generally have typically crustal origin such as Fe, Al and Ti are chosen as reference metal. Ti was not measured in this study and Fe behaves differently in oxic and anoxic environment, [8]. Hence, we choose a conservative element with high crustal abundance such as Al, as the reference metal in this study [25, 26].

2.4.3 Marine and non-marine sources and marine enrichment factor

Ionic composition of rainwater helps to understand the relative contribution of marine and non-marine sources. Generally, Na is taken as reference element with the assumption that all Na is of marine origin. The equation for the non-sea salt contribution can be written as \(\left[ {NSS - X} \right]_{I} = \left[ {X_{i} } \right] - [Na\, + ]_{i} \left[ {\frac{{\left[ X \right]}}{{\left[ {Na + } \right]}}} \right]_{{sea{\text{ }}salt}}\)

[NSS-X]i is the concentration of non-sea salt concentration of species X in sample i, [Xi] is the total measured concentration of chemical species X in sample i, [Na+]i is the concentration of Na+ in sample i, and {[X]/[Na+]}sea salt is ratio of these species measured in sea water [66]. Furthermore, marine enrichment factors of the anions with respect to Na are estimated as follows: \(EF =\)[X/Na+] rain/ [X/Na+] seawater where X is the ion of interest.

2.4.4 Wet deposition flux (WDF)

The volume-weighted mean concentrations (VWM) of trace metals (mg L−1) and wet deposition fluxes (mgm−2) were determined using Eqs. (1) and (2), respectively:

$$C_{{VWM}} = ~\mathop \sum \limits_{{i = 1}}^{n} \left( {Ci~ \times Pi} \right)~/\;~\mathop \sum \limits_{{i = 1}}^{n} Pi$$
(1)
$$WDF\, = \,C_{{VWM}} \times P_{t} /1000$$
(2)

where CVWM is the concentration of VWM, Ci (mg L−1) are the concentrations of the measured TE, Pi (mm) are the rainfall amount of individual precipitation event, and WD is the wet deposition.

WDF refers to the monthly/seasonal/annual wet deposition flux expressed in mgm−2, which is calculated by multiplying the CVWM (mg L−1) by the volume of monthly/seasonal/annual rainfall amount in mm. For the present study, seasonal Pt in mm will be for the months of July to October 2019, i.e. those during which the samples were collected and annual Pt for the entire year [18, 127]. Precipitation data were obtained from archive of Regional Meteorological Centre Kolkata, via the online portal at http://imdkolkata.gov.in/.

3 Results and discussion

3.1 PM10 concentration and trace metal composition

The PM10 concentrations ranged between 42 and 69 µg/m3 (average 56 ± 9 µg/m3, 1SD) during the sampling period which is well within the National Ambient Air Quality Standards (NAAQS) [121] prescribed value of 100 µg/m3 for 24 h.

The overall trend of the average elemental composition of PM10 is (Na, Al, Ca, Fe and Mg) > 100 ng/m3, (Zn, Mn and Pb) > 10 ng/m3, and (Ni, Cu, V, Cr and Cd) < 5 ng/m3 (Fig. 2, Table 1). Concentrations reported in this study are less than those reported in previous studies from Kolkata [25, 57]. Generally, the pollution levels have a seasonality and are minimum during the monsoon due to washout effect. The previous studies reported trace elements in PM during winter [25] when pollution is highest due to thermal inversion and annual average [57], which is generally higher than monsoon minima. Hence, absolute concentrations are inadequate to understand the sources, and thus we use the crustal enrichment factor (EF).

Fig. 2
figure 2

Box plot of metal concentrations in PM10 (red box) and rainwater (blue box) from Kolkata

Full size image
Table 1 Summary statistics of major ions and trace metals in rainwater (n = 13) and PM10 (n = 10) measured during the study period
Full size table

As per the EF calculations, any element with natural origin should have EF ~ 1. However, to account for the local variability of upper continental crustal composition which forms the immediate precursor of the aerosol metals, EF ≤ 10 is, considered to have a natural origin [25,26,27, 57, 59]. In PM10, Na, Mg, Al, Ca, V, Cr, Mn and Fe have EF < 10 indicating natural origin. Ni and Cu have EF between 10 and 100 indicating moderate enriched by anthropogenic sources. Very high EF (EF > 100) of Pb, Cd and Zn are most likely associated with industrial, high-temperature combustion and vehicular emissions (Fig. 3b). The general trend of EFs observed in our study is consistent with those reported from Kolkata by [25] indicating the presence and similar intensity of anthropogenic sources. Though EF can differentiate between crustal and anthropogenic origins, source apportionment of PM is difficult to achieve solely based on EF and elemental concentrations. Thus, principal component analysis (PCA) and elemental ratios are used to better understand the anthropogenic sources of the metals.

Fig. 3
figure 3

A) Concentration trends of the measured elements in PM10 and rainwater. B) Enrichment factors (EF) of metals, calculated using Al as reference. EF is calculated with respect to average upper continental crust (UCC) composition from [93]. Metals with values of EFs < 10 have crustal sources, and those with values of EFs > 10 indicate anthropogenic influence

Full size image

As the objective of the study is to understand sources of trace metals in rainwater, we exclude the major ions in rainwater such as Na, Ca and Mg from PCA analysis in both the PM10 and rainwater samples. Four factors can explain 92% of the total variation in the PM10 metal dataset (Table S2, Fig. 4). Communalities of all metals ranged between 0.85 and 0.99 which is an indication of satisfactory apportionment of each element to the identified factors. Physical interpretation of each factor or source was based on its association with strong loading of marker elements, typically emitted from that source.

Fig. 4
figure 4

Principal component analysis (PCA) loading three-dimensional plot for 10 trace elements in aerosol (PM10) and rainwater

Full size image

The first factor is associated with Al, Fe, Mn, and Cr and can explain 41.8% of the total variance. Al and Fe are typically crustal metals, and all the metals have EF < 10. Hence factor 1 represents crustal source [5, 21, 113].

Factor, associated with Ni, Cd and Zn explains 27.8% of the total variance. The largest sources of atmospheric Ni are coal and liquid fuels consumption [118]. Other minor sources include non-ferrous metal smelting, cement production, brick production, glass manufacturing. The ferrous and non-ferrous metal industries and fuel combustion are estimated to account for the largest emissions of atmospheric Cd [52, 85]. Cd has the highest EF and can also originate from the production and recycling of Ni–Cd batteries [117]. Major anthropogenic sources of Zn are coal combustion and industrial activities [75, 84]. Hence factor 2 represents coal combustion and industrial emissions.

The third factor, with marker elements Pb and V, explains 13% of the total variance. Both Pb and V are characteristics of road traffic or vehicular emissions [5, 42, 71]. Though leaded gasoline was completely phased out from India in the year 2000, present-day diesel and unleaded gasoline still contain trace amounts of Pb [22]. The large fleet of vehicles on city roads magnify Pb emission to the atmosphere. V/Pb ratios can differentiate vehicular sources from coal combustion and industrial emissions. Coal is generally enriched in V (mg/g level) resulting in V/Pb ratio of > 1, whereas non-coal burning industrial sources have V/Pb ratio <  < 1 [40]. V/Pb ratio of vehicle exhaust is in between and is usually < 1. The Kolkata aerosols have average V/Pb ratios of 0.30 ± 0.20, 1SD (range 0.08–0.59), indicative of vehicle exhaust emission.

Finally, the fourth factor, primarily loaded with Cu, explains 9.4% of the total variance. Fossil fuel combustion, commercial metallurgical processes and waste incineration are the major sources of atmospheric Cu emission [63].

3.2 pH and conductivity of rainwater

The pH of the rainwater from this study ranged from 5.4 to 8.0 (average 6.8 ± 0.74, 1SD). Out of 13 rainwater samples, approximately 15% were acidic (pH < 5.61), 15% were slightly acidic (pH 5.6–6.5) and 54% of the samples were alkaline with pH > 7. The pH of natural precipitation in pristine environment is 5.6 and is controlled by interaction between water droplets and atmospheric CO2 [19]. However, in polluted settings, the pH of rainwater is mainly regulated by the abundance of acids forming chemical species such as SO42− and NO3−, and basic gas such as NH3. Soil dust containing Ca2+, Mg2+ and K+ can balance acidification, due to their buffering capacity [60, 72, 99, 101, 110, 120]. The average pH of the present study (pH = 6.8) is slightly higher than the Indian average of 6.5 [17]. The pH range of the present study is comparable with those reported in earlier studies from Kolkata [17] but less alkaline than those reported by [68] (Table 2). The sulphate ion showed strong negative correlation with pH (r2 =  − 0.8) but not the nitrate. The almost neutral pH of the precipitation is due to the neutralization caused by different cations. When the sum of ([Ca2+] + [Mg2+]) is fitted against [SO42−], the correlation coefficient is 0.73, indicating that Ca and Mg carbonate is the major neutralizing base [5]. Slightly alkaline pH has been reported from many other cities and states of India as well as places around Asia (Table 2). In general, Indian precipitation has higher pH than natural rainwater resulting from high loading of alkaline rich atmospheric particulate matter though acid rain has been reported from coal mining city of Dhanbad and the national capital of Delhi due to the influence of SO42− ions [91, 109].

Table 2 Comparison of pH of rainwater of the sampling site in Kolkata with some other selected sites of India and the world
Full size table

Conductivity in the present study ranged from 12.6 to 188 μS/cm (average 55 ± 46 μS/cm, 1SD). Low conductivity in precipitation is an indicator of good environmental quality in the atmosphere [125]. The common conductivity range of precipitation is 5–1000 μS/cm [4]. There were two previous studies from Kolkata. One reported similar conductivity (47.87 μS/cm) [17], and the other study reported much higher conductivity (370 ± 297 μS/cm) that was attributed to industrial emissions of cations and anions in the megacity atmosphere [68]. The conductivity of the present study is compared to other Indian cities given in Table 2. Precipitation conductivity is attributed primarily to the total soluble ionic components present in the rainwater hence reflecting the impact of atmospheric particulate matter on rainwater chemistry [80]. In this study, the conductivity strongly correlates (r2 = 0.85) with Ca most likely due to presence of CaCO3 as a significant solute [3].

3.3 Chemical composition of rainwater

Inorganic composition of rainwater comprises the major anions (Cl−, NO3−, SO42−), NH4+ from anthropogenic emissions, major elements (Ca, Mg, Na, K) and trace metals. In this study, K and NH4+ were not measured. Concentrations in rainwater follow the order of (Na and Ca) > 1000 µg/L, Mg > 100 µg/L, (Al, Fe and Zn) > 10 µg/L and (Mn, Ni, Cu, V, Cr, Cd and Pb) < 10 µg/L (Fig. 2).

The marine and non-marine contributions of the anions Cl− and SO42− were determined. The ratios of Cl− and SO42− with respect to Na+ in sea water are 1.16 and 0.125, respectively [66]. Average Cl−/Na+ ratio of rainwater samples (1.12 ± 0.17) is similar to sea water ratio of 1.16. Significant correlation between Cl− and Na+ (r = 0.96) confirmed sea salt origin. Average SO42− /Na+ ratio of rainwater samples (1.132 ± 0.90) is higher than sea water. Calculations show sea salt fraction of SO4 2− is only 12%. Furthermore, marine enrichment factor of Cl− is 1 indicating marine source and that of SO42− is 9 indicating contribution of non-marine sources. The non-sea salt fraction of SO42− may be sourced from soil dust containing gypsum, pyrite or may be derived from anthropogenic processes, primarily combustion [10, 35, 65, 83, 112].

Removal of aerosols by wet deposition occurs primarily by two mechanisms, below-cloud scavenging and within-cloud scavenging [43]. For below cloud scavenging, trace metal composition of the ambient aerosol controls the composition of rainwater. Relative abundance of metals in rainwater depends on emission sources and solubility of metals. The general concentration trend of the major elements in rainwater followed the aerosol Na, Ca, Mg trend (Fig. 3a). The overall concentration trend of the metals Cu, Zn, Cd, V, Cr and Pb follows the trend in PM10 indicating at least part of the metals are derived from aerosol leaching (Fig. 3a). However, Al and Fe do not follow the trend indicating limited leaching of these two elements from aerosol. Natural sources of Al and Fe in aerosol are crustal dust, and these two elements are mostly locked inside aluminosilicate structures [93] that are difficult to leach in rainwater. The sampling location is on the alluvium of the Indo Gangetic Plain (IGP). The common aluminosilicates that are found in the IGP soil are illite/mica, smectite, kaolinite and feldspars. The EFs of trace elements in rainwater of Kolkata are shown in Fig. 3b and compared with the EF of PM10 trace elements. Out of all the trace elements measured, Fe is the only non-enriched element with EF < 10. The major element Ca and Mg in rainwater have high EF indicating anthropogenic influences. In cities, the sources of Ca and Mg can be cement from construction dust and road dust because of the abundance of alkaline materials in road coverings [37]. V, Mn, and Cr are moderately enriched with EF between 10 and 100. Cu, Ni, Zn, Pb and Cd were highly enriched with EF > 100 indicating strong anthropogenic sources.

Cu, Ni, Zn, Pb and Cd have EF > 10 in PM10 collected in between rain events and were identified to have anthropogenic sources related to non-ferrous metal industry, waste incineration and coal combustion [63, 84, 85]. V, Mn, Co and Cr in aerosol have EF between 4 and 9. However, their EF in rainwater increases to greater than 10 but less than 100 in rainwater.

Figure 4 presents the three dissolved trace metal clusters for rainwater during monsoon period. Varimax-rotated PCA identified three sources of trace metals in the rainwater and can explain 79% of the variance in the data (Table S3). The first factor explains 38.6% variance and is loaded with crustal element such as Al, and the rest are anthropogenic metals such as Mn, V, Cd and Pb. Factor 1 can be attributed to vehicular emission and road dust [107]. Factor 2 accounts for 27% of the variance and has strong loadings of Fe, Zn, Cu and Cr and is possibly soil dust mixed with industrial emission. Factor 3 explains 13.7% variance and has strong loading of Ni and moderate loading of Zn and is possibly related to coal combustion emission [84].

VWM and average concentrations of trace elements in our study were compared with those reported from India and worldwide (Table 3). Selected worldwide stations are either large urban centres and/or stations close to the coast like Kolkata. A remote station in Lhasa, Tibet, is also chosen for comparison. Trace elements in two successive rain events from the Indian city of Lucknow reported higher concentrations of Cu, Zn, Cr, Cd and Pb. This is possibly because concentrations of these elements are several times higher in Lucknow PM10 compared to Kolkata during monsoon [103]. However, the trace element concentrations in the Kolkata rainwater are in the same order of magnitude with those reported from global rainfall data over urban centres. The concentrations of Ca were higher than those reported from both the stations in Brazil and Europe. The high concentration of Ca in Kolkata rain is primarily due to the influence of soil dust and endless constructions which contains large fractions of CaCO3. The concentration of marine elements such as Na and Mg in rainwater of Kolkata is higher than most of the compared cities but lower than that of Coastal Station in Western Europe and comparable with Northern & Southern Jordan [3, 5, 28]. The Bay of Bengal coast is approximately 125 km south of Kolkata and presumably influences the rainwater composition. Concentrations of crustal elements such as Al and Fe measured in Kolkata precipitation are low relative to other sites. Anthropogenic elements such as Zn and Pb were significantly lower than Kolkata rainwater compared to cities in Greece, Iran, Jordan, China and Lucknow. Pb shows large variation in the global dataset. In Jordan, high concentrations of Pb are attributed to biomass burning and industrial activities [3, 5]. High Pb was also reported from China and Singapore. While Pb in China was due to industrial emission, high Pb in Singapore rain was due to atmospheric scavenging of gasoline Pb released from neighbouring Indonesia. Interestingly, though India is the second largest emitter of atmospheric Pb in Asia, after China [70], Pb concentration in rainwater is < 1 ppb indicating low solubility of atmospheric Pb over Kolkata and hence limited bioavailability of the toxic metal. Most of the metals (except the crustal elements Al and Fe) in Kolkata are higher than in Tibetan rainfall as expected, but interestingly, Pb concentrations are comparable and only ~ 0.5 times higher in Kolkata rain.

Table 3 The volume-weighted mean (VWM) concentrations of major ions and trace metals (µg/L) in this study are compared to those obtained from other parts of the world
Full size table

3.4 Comparison of precipitation chemistry with groundwater and surface water

With increasing population pressure, there is growing reliance on groundwater for farming and drinking purposes. Human activity has threatened freshwater reserve by polluting and depleting groundwater reserve by overdrawing [123]. Rainwater is relatively free from impurities and may be one of the alternative sources of drinking water.

The drinking water standards of India consist of several criteria such as: (1) perceptible parameters and inorganic constituents, (2) organic substances, (3) pesticides, (4) disinfectant and disinfectant by-products, (5) radioactive constituents, and (6) micro-organisms [46]. The present study focuses only on the inorganic components of the rainwater. In Table 4, we have compared rainwater chemistry with other sources of drinking water such as untreated groundwater and river water from India. Kolkata rainwater meets the Indian and the WHO standards for drinking water, while the concentrations of toxic metals such as Pb and Ni of river and groundwater are higher than the guideline values. Pb is long known to affect physical or mental development in infant and children and causes high blood pressure and kidney problems in adults [15, 20]. Ni is a well-known human carcinogen that affects the activity of α-tocopherol, the most common lipid soluble antioxidant in human body [58]. The concentration of major ions such as Na, Ca and Mg is higher in river and groundwater compared to rainwater, but they are within the limits prescribed by WHO and Indian standards. Fe exceeds safe limits for drinking water in the ground and surface water. All the measured anions (chloride, sulphate and nitrate) are significantly higher in the ground and surface water as compared to the rainwater and exceeds the WHO and Indian Standards. Nitrate in drinking water is often reduced to nitrite in the stomach by gastric acids and with subsequent reactions with amines and amides results in the formation of N-nitroso compounds that have been linked to different types of cancer [30]. Nitrate (12.5 ± 11.8 ppm) in the rainwater meets the Indian, WHO and EU standards, but fails to meet the US EPA standard of 10 ppm. Therefore, rainwater needs a bare minimum treatment with respect to inorganic constituents before use as opposed to groundwater and river water, and this makes the use of rainwater more convincing as a drinking source.

Table 4 Comparative table showing rainwater, river and groundwater and water quality standards. Units of trace elements are in ppb (µg/L) and those of the anions are in ppm (mg/L)
Full size table

3.5 Wet deposition flux (WDF) of the trace metals

The WDF calculated during the current study are shown in Table 5 and compared with the WDF data in different Asian countries from the literature. The atmosphere includes a variety of nutrient ions and trace metals adsorbed onto particulate matter that are scavenged by rain. Some of these ions and metals are plant nutrients (such as nitrate, ammonium, K, Ca, Mg, Fe) and some of them are toxic such as Cr, Cd, Pb, which affect the physico-chemical properties of the soil [32, 48].

Table 5 The wet deposition fluxes (WDF) in mg/m2/year from this study compared with those reported for other locations in Asia. As sampling was conducted from July to October 2019, WDF for those four months are also reported for comparison
Full size table

WDF depend on the solubility of the trace metals and particle size distribution [86]. In Kolkata, the magnitude of the concentrations and fluxes of the different elements were in the order of: Ca > Na > Mg > Al > Zn > Fe > Mn > Cu > Ni > Pb > Cr > V > Cd. Of the primary crustal elements, Ca exhibited the highest flux of 242 mg/m2/yr followed by Na, Mg, Al and Fe. Al, Fe and Mn represent higher loadings because of their high concentrations in dust derived from crustal materials [23] and Na has a sea salt origin and comparable with the WDF over IGP. Higher flux values of Ca and Mg over IGP compared to Kolkata are due to enhanced anthropogenic emissions from power plants, refineries, stone crushing units and numerous medium and small-scale industries over IGP [119]. Among the anthropogenic metals, Zn had the highest WDF, followed by Cu, Ni, Pb, Cr, V and Cd. Despite differences in sampling periods reported in previous studies and the probability of inter-annual variability, WDFs of elements like Cd, V, Cr, Ni and Cu are in good agreement with those recorded from other sites. The deposition fluxes of Ni, Cu and Zn at Kolkata were more or less similar to Izmir, Turkey, and Singapore both of which are large coastal urban centres where the sources are attributable to local emissions from anthropogenic activities [47, 102]. V and Cr fluxes of this study are comparable with different cities of China and Japan but much less than those reported from Turkey and Singapore. Leather tanning, textiles industries and electroplating are significant anthropogenic sources of Cr in soil, whereas fossil fuel (both coal and oil combustion) and waste incineration releases Cr in the atmosphere [50]. V is released from combustion of fossil fuels. Wet deposition of Cd is comparable with global data. Cd is primarily released from coal combustion, vehicle emissions and industries. Pb flux is lower than all the compared sites. Post-phasing out of leaded gasoline the largest anthropogenic sources of Pb is coal combustion followed by nonferrous metal smelters and vehicle exhaust, besides the natural background from regional and local soil. Among the compared sites, Mt. Qomolangma in China which is regarded as a remote background site has the least WDF for almost all the trace elements.

3.6 Limitations of the study

The current study collected rainwater and aerosol in between rain events during the monsoon season only. In 2019, the total rainfall over Kolkata was 1642 mm out of which 1127 mm or 69% occurred during the sampled monsoon months between July and October. The remaining months of the year except November received less than 100 mm rainfall. Hence, seasonality of rainwater chemical composition could not be measured in this pilot study. Furthermore ammonia (NH4+) in rainwater was not measured in this study; thus, neutralization factors could not be analysed. Future study plans to focus on seasonal variation of rainwater metal and major ions concentrations in urban centres and compare it with a less polluted rural site to better understand sources of pollutants in rainwater.

Access to safe drinking water remains an urgent requirement in India, as a large fraction of Indian households (30% of urban and 90% of rural households) are still dependent on either groundwater or untreated surface water [2]. Though access to drinking water has improved in India over the last decade, the enormous adverse health effects of contaminated water continue. Generally, contaminants in harvested rainwater are derived primarily from the atmosphere; however, rainwater quality deterioration may occur during harvesting, storage and household use [77]. To correctly access the treatment needs of harvested rainwater for drinking purpose, both harvesting conditions and rainwater quality with respect to chemical (organic and inorganic constituents, pesticides, radioactive components, disinfectants), physical and microbiological components should be assessed [1]. Identifying pollutants that are incorporated from the harvesting process and organic and microbiological pollutants quantification of rainwater was not within the scope of this study.

4 Conclusions

The present work is a pilot study to assess metal concentrations and major anions in rainwater and quantifying wet deposition flux of the trace metals from a polluted megacity in India. Trace metal chemistry of rainwater is gaining importance as rain water harvesting is one of the strategies that the water sector is increasingly adopting to cope with future climate change. In many areas in China and Middle East, Pb exceeds WHO guideline limits for safe drinking water. In Lucknow, India, several toxic trace metals such as Pb, Cr and Cd exceeds Indian Standards. Kolkata in spite of having poor air quality experienced rainfall with all the trace elements and anions within WHO guideline limits. Hence the rainwater is deemed suitable for drinking purpose with respect to metal content and anions and does not require pre-treatment to remove toxic metals. Comparison of concentration trends and crustal enrichment factor for the trace metals of PM10 and rainwater shows preferential leaching of metals such as Cu, Zn, V, Cd and Pb. A soluble fraction of the metals enter the ecosystem through wet deposition. The annual WDF for Pb was much less when compared to other sites around the globe. However, WDF of the remaining metals are comparable to the global data.