**1. Introduction**

Many atmospheric organic and inorganic compounds have a tendency toward dissolution in atmospheric water (cloud droplets and rainwater) [1]. Rainwater is very sensitive to changes in the chemistry of the atmosphere [2]. The physico-chemical properties of the precipitation are influenced by local sources (natural and anthropogenic) [1] as well as long-range transport of dust and aerosols [3,4]. It was found that 40% of the dissolved organic carbon (DOC) in the rainwater is resistant to bacterial degradation and may be transported by long-distance in the atmosphere before being removed by rain [5]. Previous studies on the rainwater chemistry in Croatia have shown that the rainwater composition is profoundly affected by the transport of the eolian dust from North Africa (very often in spring and summer) as well as polluted aerosols from the other Europe countries [6,7]. Atmospheric

transport and deposition is an important pathway of various substances, including organic compounds and trace metals to the surface environment [5,8–13]. Removal of substances from the atmosphere by precipitation affects chemistry, biogeochemistry, and sedimentation in the aquatic environment [11,14]. These processes have a decisive role in global changes and in environmental protection [5,8–11].

Dissolved organic matter (DOM), measured as dissolved organic carbon (DOC), is an important component of the rainwater [5]. Removal of the atmospheric DOC by rainwater, including non-oxidized organic compounds, decreases the generation of carbon dioxide [15]. This process is crucial for the evaluation of the global carbon cycle [5,15]. The climate change in recent decades has led to increased interest in the study of DOC in atmospheric precipitation [16–20] and surface waters [14] (and references therein).

Surface-active substances (SAS) make up a portion of rainwater DOC [21–23] (and references therein) and the water-soluble organic carbon (WSOC) fraction of atmospheric aerosols [24] (and references therein). These compounds are the most reactive part of DOC [25] because they tend to adsorb on the natural phase boundaries (air-water, water-sediment/suspended particles) and, in this way, significantly affect the transfer of mass and energy. In our previous work, we have shown that some metal ions can accumulate in adsorbed layers of organic molecules [23]. This mechanism could be important for the geochemical cycle of metals.

Film-forming compounds can influence the microphysical processes of humid aerosol particles and atmospheric droplets [24,26]. Because of these properties, atmospheric SAS plays an essential role in various atmospheric processes that reflect climate change [24] (and references therein). Physico-chemical properties of OM are often more important than the amount present [22]. Atmospheric humic-like substances, HULIS were found to be an important class of SAS in aerosol particles [24,27,28] (and references therein). Rainwater SAS are compositionally different from those found in surface waters [29]. Atmospheric HULIS decrease the surface tension more efficiently than the terrestrial and aquatic humic substances because of the different composition [27]. Some studies also suggested that atmospheric SAS could be "biosurfactants" of microbial origin [24]. Quantification and characterization of the SAS in the atmospheric deposition are important for a better understanding of its contribution to organic carbon (OC) cycling.

Complexation of OC with metal ions has a significant impact on mobility, solubility, and bioavailability of metals in precipitation and further in the aquatic environment [30–36].

The organically complexed metals such as Cu, Co, Ni, Pb, Cd, and Zn have been observed in urban rainwater [32] (and references therein). Metal complexing capacity (MeCC) refers to the amount of OM present that can specifically bind metal ions, and it is often used as a parameter for qualitative and quantitative characterization of the OM in natural waters [32–35]. These investigations are usually focused on the complexation of Cu ligands [32–35] because the Cu ions have a strong affinity for naturally occurring organic ligands [32,33]. In the urban atmosphere, Cu is derived from fossil fuel combustion, exhaust emissions, and industrial processes [37] (and references therein), although the significant atmospheric Cu in western Europe comes from brake wear [38].

In the atmosphere, sulfur is mainly emitted in its reduced form [39]. Carbon disulfide (CS2), dimethyl sulfide (CH3SCH3, DMS), and hydrogen sulfide (H2S) are the most frequently reduced sulfur species (RSS), which are found in the atmosphere [40]. In Orlovi´c-Leko et al. [41], inorganic or organic RSS has been confirmed at nano-levels in rainwater samples by electrochemical methods. The chemistry of RSS is crucial for many environmental problems, such as acid rain and climate change [39]. While sulfur species, together with SAS, could play a crucial role in speciation and solubility of trace metals in aquatic systems [42].

In an urban environment, metals are emitted primarily by anthropogenic sources, which include traffic, industry, corrosion of construction materials, and waste incineration [43].

An important ecological property of metals is their persistence and ability to accumulate in ecosystems [43]. Measurements of TEs concentrations in the rainwater could be critical to understanding of their geochemical cycling [44–47]. The investigation of the soluble fraction of metal in atmospheric deposition is a priority in many studies [36,47–49] because the solubility of TEs

affects their bioavailability. The solubility of TEs depends on the pH values of rainwater and the size of particles [48]. The particles of anthropogenic origin are highly soluble [46] (and references therein). of particles [48]. The particles of anthropogenic origin are highly soluble [46] (and references therein).

affects their bioavailability. The solubility of TEs depends on the pH values of rainwater and the size

The aim of this study is the characterization of OM (through measurement of DOC, particulate organic carbon POC, SAS, RSS) and TEs in precipitation of an urban environment, in Zagreb city, Croatia. In our previous studies, conducted in 1998–1999 and 2003–2007, spatial and temporal variability of DOC concentrations in bulk precipitation of urban and coastal areas in Croatia have been discussed [21,22]. However, in accordance with the literature that emphasized a lack of data on DOC measurements in rainwater in central and eastern Europe [17], as well as trends based on the long-term observations of DOC in rainwater [8] (and references therein), in this paper, we focus on studying the possible changes that may have occurred in the concentration and reactivity of DOC between two sampling periods (1998–1999 and 2009–2011), in line with socio-economic and global changes. Namely, Croatia ratified the Kyoto Protocol (April 2007), thereby committing to reducing emissions of greenhouse gases in the period 2008–2012 compared to the baseline emissions in nineties (1990). Also, as a prerequisite for achievements of EU Directive 2006/32/EC on energy efficiency and energy services, Croatia prepared a program "Adjustment and upgrades of energy development strategies," which foresees adaptation of the power plant in Zagreb to use renewable energy. In the studied sampling period the Zagreb thermal power plant was operated on fuel oil, which may affect DOC and TE concentrations. Therefore, this study on the rainwater OM and TE is unique for Zagreb, and Croatia, and could serve as a baseline for future investigations in the area. The aim of this study is the characterization of OM (through measurement of DOC, particulate organic carbon POC, SAS, RSS) and TEs in precipitation of an urban environment, in Zagreb city, Croatia. In our previous studies, conducted in 1998–1999 and 2003–2007, spatial and temporal variability of DOC concentrations in bulk precipitation of urban and coastal areas in Croatia have been discussed [21,22]. However, in accordance with the literature that emphasized a lack of data on DOC measurements in rainwater in central and eastern Europe [17], as well as trends based on the long-term observations of DOC in rainwater [8] (and references therein), in this paper, we focus on studying the possible changes that may have occurred in the concentration and reactivity of DOC between two sampling periods (1998–1999 and 2009–2011), in line with socio-economic and global changes. Namely, Croatia ratified the Kyoto Protocol (April 2007), thereby committing to reducing emissions of greenhouse gases in the period 2008–2012 compared to the baseline emissions in nineties (1990). Also, as a prerequisite for achievements of EU Directive 2006/32/EC on energy efficiency and energy services, Croatia prepared a program "Adjustment and upgrades of energy development strategies," which foresees adaptation of the power plant in Zagreb to use renewable energy. In the studied sampling period the Zagreb thermal power plant was operated on fuel oil, which may affect DOC and TE concentrations. Therefore, this study on the rainwater OM and TE is unique for Zagreb, and Croatia, and could serve as a baseline for future investigations in the area.

Data treatment for TEs included the calculation of their daily fluxes and percentage of solubility. In order to study the possible sources of TEs, correlation analyses, and enrichment factors (EFs) were considered. Data treatment for TEs included the calculation of their daily fluxes and percentage of solubility. In order to study the possible sources of TEs, correlation analyses, and enrichment factors (EFs) were considered.

### **2. Experiments 2. Experiments**

#### *2.1. Area Description 2.1. Area Description*

The sampling site (Figure 1) was located in the center of Zagreb (45◦48024" N and 15◦57050" E), on the roof (20 m above ground level) of the Faculty of Mining, Geology, and Petroleum Engineering building. The site is not surrounded by rows of buildings, and therefore, airflow was not channeled, obstructed, or restricted. It is close to the parking lot and two main roads characterized by heavy traffic all day long. The thermal power station (burns gas and oil fuel) is located approximately 2 km from our sampling location. However, it is important to note that there are no large industries in Zagreb and that the topography of the sampling site did not change over the course of the sampling period. The primary local pollution sources are traffic and two thermal power stations. Gas is a standard fuel for domestic heating. The sampling site (Figure 1) was located in the center of Zagreb (45°48′24″ N and 15°57′50″ E), on the roof (20 m above ground level) of the Faculty of Mining, Geology, and Petroleum Engineering building. The site is not surrounded by rows of buildings, and therefore, airflow was not channeled, obstructed, or restricted. It is close to the parking lot and two main roads characterized by heavy traffic all day long. The thermal power station (burns gas and oil fuel) is located approximately 2 km from our sampling location. However, it is important to note that there are no large industries in Zagreb and that the topography of the sampling site did not change over the course of the sampling period. The primary local pollution sources are traffic and two thermal power stations. Gas is a standard fuel for domestic heating.

**Figure 1. Figure 1.**Position of the sampling site (adapted from Google Maps). Position of the sampling site (adapted from Google Maps).

Zagreb has a continental precipitation regime, with its maximum in the warm months of the year and a secondary maximum in autumn. The analysis of annual precipitation amounts from 1961 to 1990 showed that in Croatia, the annual precipitation amount in 2009 (794.8 mm) and 2011 (520.8 mm) was below the average, and in 2010 (1155.1 mm) it was above the average [50–52]. A comparison of monthly precipitation in 2009/2010 in the Zagreb area is presented in Figure 2 [50,51]. During the study period, daily precipitation ranged from 0.3 to 28.8 mm (Croatian Meteorological and Hydrological Service Data, DHZM). Zagreb has a continental precipitation regime, with its maximum in the warm months of the year and a secondary maximum in autumn. The analysis of annual precipitation amounts from 1961 to 1990 showed that in Croatia, the annual precipitation amount in 2009 (794.8 mm) and 2011 (520.8 mm) was below the average, and in 2010 (1155.1 mm) it was above the average [50–52]. A comparison of monthly precipitation in 2009/2010 in the Zagreb area is presented in Figure 2 [50,51]. During the study period, daily precipitation ranged from 0.3 to 28.8 mm (Croatian Meteorological and Hydrological Service Data, DHZM).

*Atmosphere* **2020**, *11*, x FOR PEER REVIEW 4 of 19

**Figure 2.** Monthly precipitation 2009/2010 at Zagreb area [50,51]. **Figure 2.** Monthly precipitation 2009/2010 at Zagreb area [50,51].

### *2.2. Sampling Method and Sample Treatment Prior Analyses 2.2. Sampling Method and Sample Treatment Prior Analyses*

Daily, bulk precipitation (27 samples) were sampled from January 2009 to May 2010. In 2011, only four samples were collected, each in different seasons. The sampling was performed using two samplers consisting of a glass funnel (25 cm diameter) and a glass bottle (2.5 L), which were installed 1.0 m above the ground for bulk rainwater sampling. Both components of the sampler were prepared according to the standard cleaning procedure. They were washed with chrome-sulphuric acid (in-house), rinsed by HNO3 (p.a. Kemika, Zagreb,Croatia), and several times with Milli-Q water (Milli-Q, 18.2 MW, total organic carbon (TOC) < 3 ppb) before being set up at the sampling location. The collected precipitation volumes ranged from 0.1 to 2.2 L (collected samples with a volume of less than 0.1 L were not considered). Daily, bulk precipitation (27 samples) were sampled from January 2009 to May 2010. In 2011, only four samples were collected, each in different seasons. The sampling was performed using two samplers consisting of a glass funnel (25 cm diameter) and a glass bottle (2.5 L), which were installed 1.0 m above the ground for bulk rainwater sampling. Both components of the sampler were prepared according to the standard cleaning procedure. They were washed with chrome-sulphuric acid (in-house), rinsed by HNO<sup>3</sup> (p.a. Kemika, Zagreb, Croatia), and several times with Milli-Q water (Milli-Q, 18.2 MW, total organic carbon (TOC) < 3 ppb) before being set up at the sampling location. The collected precipitation volumes ranged from 0.1 to 2.2 L (collected samples with a volume of less than 0.1 L were not considered).

The bulk sampler (bottle/funnel) collected dry deposition of gases and particles. However, it was found that approximately 80% of carbonaceous aerosols are removed by wet deposition (via precipitation) [53]. Furthermore, according to the European Committee for Standardization (CEN), the bulk sampler can also be used for measurements of atmospheric deposition of metals in industrial and urban sites, mainly in the case of daily samples [6,47] (and reference therein). The bulk sampler (bottle/funnel) collected dry deposition of gases and particles. However, it was found that approximately 80% of carbonaceous aerosols are removed by wet deposition (via precipitation) [53]. Furthermore, according to the European Committee for Standardization (CEN), the bulk sampler can also be used for measurements of atmospheric deposition of metals in industrial and urban sites, mainly in the case of daily samples [6,47] (and reference therein).

Volumes and the pH values of samples were measured just after collection. For measurements of TEs in the filtered fraction, the samples were filtered by use of the Millipore vacuum filtration system through a 0.45 µm cellulose-nitrate membrane filter (Sartorius). For DOC measurement, the filtered fraction was obtained by the use of the same filtration system with Whatman glass fiber filters, pore size 0.7 µm (Whatman, Grade GF/F, d = 47 mm). After preservation with the HgCl2 solution, it was stored in a cold dark place until analysis. Particulate organic carbon (POC), defined as organic matter larger than 0.7 µm, remained on the filter. Parameters, SAS, CuCC, and RSS, were measured in the original samples within two days. Prior to analysis, the samples were stored at 4 °C. Volumes and the pH values of samples were measured just after collection. For measurements of TEs in the filtered fraction, the samples were filtered by use of the Millipore vacuum filtration system through a 0.45 µm cellulose-nitrate membrane filter (Sartorius). For DOC measurement, the filtered fraction was obtained by the use of the same filtration system with Whatman glass fiber filters, pore size 0.7 µm (Whatman, Grade GF/F, d = 47 mm). After preservation with the HgCl<sup>2</sup> solution, it was stored in a cold dark place until analysis. Particulate organic carbon (POC), defined as organic matter larger than 0.7 µm, remained on the filter. Parameters, SAS, CuCC, and RSS, were measured in the original samples within two days. Prior to analysis, the samples were stored at 4 ◦C.

### *2.3. Chemical Analyses 2.3. Chemical Analyses*

### 2.3.1. Organic Matter (DOC, POC Measurements)

2.3.1. Organic Matter (DOC, POC Measurements) DOC and POC content was determined by the high-temperature catalytic oxidation (HTCO) method at a TOC-VCPH instrument (Shimadzu, Japan). The DOC and POC concentrations of each DOC and POC content was determined by the high-temperature catalytic oxidation (HTCO) method at a TOC-VCPH instrument (Shimadzu, Japan). The DOC and POC concentrations of each

sample were calculated as an average of three replicates. The limits of quantification (LOQ) are 0.228

sample were calculated as an average of three replicates. The limits of quantification (LOQ) are 0.228 mg/L for DOC and 7.11 µg/L for POC. The precision of DOC and POC measurements, given as the relative standard deviation (RSD) was based on the analysis of selected samples and the reference materials; RSD never exceeded 6% and 5%, respectively. Both measurement procedures were validated through international intercalibrations as a prerequisite for accreditation achieved in 2017 (HRN EN ISO/IEC 17025:2007).

### 2.3.2. Surface Activity of Rainwater DOC

The surface activity of DOC was determined by the electrochemical method of alternating current voltammetry with out-of-phase mode by using a µ-Autolab (Electrochemical Instrument Eco Chemie, Metrohm Autolab B.V., Utrecht, The Netherlands) Potentiostat connected with 663 VA Stand Metrohm mercury electrode [54]. The concentration of surface-active substances (SAS) was expressed as equivalent in mg/L to a model substance, the nonionic surfactant polyoxy ethylene-*t*-octylphenol (Triton-X-100). The detection limit of SAS determination was 0.01 mg L−<sup>1</sup> equivalent of T-X-100, with LOQ of 0.03 mg L−<sup>1</sup> . The method enables a rough characterization of SAS based on its adsorption behavior, i.e., hydrophobic—hydrophilic interactions, as seen by electrochemical measurements at the Hg electrode in the water solution [54].

### 2.3.3. CuCC Measurements

CuCC was measured by using differential pulse anodic stripping voltammetry (DPASV) by direct titration of the sample with copper ions [34]. The value of CuCC, as well as the corresponding stability constant, was calculated by applying the linear transformation plot [34] (and references therein). The RSD of the mean value calculated for five independent measurements was below 10%.
