This study was carried out in the Lake Kivu basin on the Democratic Republic of Congo side. Lake Kivu is located at 1463 m above the sea level, between 1˚34.5'S and 2˚30'S and 28˚50'E and 29˚23'E. Its basin is dominated by steep (<3%) mountainous topography underlain by sedimentary sandstone and siltstone soils [10]. The climate is tropical with bimodal rainfall distribution. Annual rainfall is about 1500 mm in the Lake Kivu basin [11]. This study considered four sampling stations: Goma in the north, Iko Island within the Lake, Lwiro in the west, and Bukavu in the south all of which are located on the DRC side of the Lake (Figure 1).

The Goma and Bukavu regions are located within urban areas in the northern and southern parts of the lake at 1˚41.58'S, 29˚22.690'E and 2˚29.817'S, 28˚51.558'E respectively. These areas are densely populated with high transport activities. Goma is also located near the active Virunga volcanoes. The Lwiro and Iko stations are situated at 2˚14.228'S, 28˚48.441'E and 2˚14.064'S, 28˚53.555'E respectively in rural areas. The dominant land use/cover in these areas is agriculture. Lwiro area is located near the Kahuzi Biega National Park Forest. These areas (Lwiro and Iko) are not densely populated.

For wet deposition, rainwater samples were collected monthly during rain events over fifteen months period (December 2017-February 2019) using polyethylene buckets (750 cm² collection area), each in four different locations in the stations

of Goma, Lwiro, Iko and Bukavu. The rain collectors were situated 1.5 m above ground level to avoid contamination with soil [12]. The start and end times of each rain event and the volume of water collected was recorded. The pH, electrical conductivity (EC) and volume measurements were conducted immediately after each sample was collected. Water samples were transported to the Laboratory of Malacology at the Centre de Recherche en Sciences Naturelles (CRSN) of Lwiro and stored at 4˚C where samples were not analysed immediately. For dry deposition, buckets containing 2 L of deionized water were exposed for 24 hours. Samples were collected at each station after 13 days (two or three times in the month). Buckets were placed on an elevated surface of 1.5 m to avoid contamination [12]. They were rinsed with several aliquots of deionized water before use. Sample water was analysed in the same manner as wet deposition samples. Before collecting samples, bird droppings that could contaminate the samples were checked, and if present, the samples were not considered for analysis. After collection, sample volume was recorded, and samples were transferred to the laboratory, where one sub-sample was filtered through a Whatman GF/F glass fiber filter (nominal pore size = 0.7 μm). The filtrate and the remaining samples were stored at 4˚C awaiting analysis. At each station (Lwiro, Goma, Iko and Bukavu), four sites were identified as representative of the land use in that area, and the result was taken as the mean for the station.

The analytical procedures used to measure major ions and other physico-chemical parameters in dry and wet deposition samples were similar to those described by other authors [13] [14] [15]. Briefly, the methods were as follows: Sodium and potassium were analysed by atomic absorption spectrophotometer (Perkin Elmer-A Analyst 400) with an air acetylene flame in the Soil Science Laboratory at Makerere University. Calcium and magnesium were determined by the Eriochrome Black T titration method using the buffer solution suggested by Golterman et al. [13]. Ammonium and nitrate (after passing through a zinc column to reduce NO 3 − to NH 4 + ) were analysed by spectrophotometric indophenol blue method [14]. Chloride was analysed by the argentometric titration method, and sulphate was analysed by the gravimetric method using barium chloride [13]. Carbonate ( CO 3 2 − ) and bicarbonate were determined by standard acid titration methods [13] [15]. The pH was measured with a digital Hanna multi-parameter probe calibrated with pH 4.0 and 10.0 buffers, and electrical conductibility (EC) was calibrated against a reference KCl solution.

Precipitation acidity originates primarily from sulfuric acid and nitric acid and neutralization by Ca²⁺ and Mg²⁺, the ratios of SO 4 2 − + 2NO 3 − /Ca²⁺ + Mg²⁺ can be considered as indicators for acidity. The ratio less than one indicates the alkaline nature of rainwater, and the ratios greater than unity indicate the presence of free anions, which are responsible for precipitation acidity [16].

An enrichment factor (EF) ratio for some ion species was calculated for samples as the ratio of the ions of interest to an ion expected to have only a soil/water origin [17]. Na⁺ is commonly taken as the best reference element for marine sources [18]. Al³⁺ and Ca²⁺ are two typical lithophilic elements normally used as reference elements for the continental crust [19]. Since there was no measurement of Al³⁺ element in dry and wet atmospheric deposition samples, Ca²⁺ was chosen as a reference element for continental origin [20].

Enrichment factors (EF) were computed as in equations 1:

EF for soil = [X/Ca²⁺]_{wet and dry deposition}/[X/Ca²⁺]_soil (1)

where: X was the concentration of the desired ion, X/Ca²⁺ ratio was from crust composition of soil in the region [21]. According to Enrichment Factors (EF), the source of major ions in atmospheric deposition can be summarized in Table 1:

To capture better the temporal variability in rates and sources of atmospheric deposition to the Lake Kivu basin, concentrations of major ions were measured two to three times a month in all four stations around the Lake Kivu, and the average was recorded.

Statistical analysis was performed using one-way analysis of variance (ANOVA), and the station means were compared using the Least Significant Difference (LSD) for the critical value p ≤ 0.05. Pearson correlation (r) test was conducted to examine potential relationships among pair-wise comparisons at each station.

The major ion dry volume weight concentration for the different stations during the study period from December 2017 to February 2019 is presented in Table 2. No significant differences were observed between stations (p = 0.98) when all the parameters were considered together, however significant differences were observed between difference dry deposition rates (p < 0.05). All the major ions

were significantly different between stations except for CO 3 2 − (p > 0.05). Major ions deposition rates varied in dry atmospheric deposition form one station to other (Table 2).

Highest Cl⁻, HCO 3 − , Mg²⁺, Na⁺ deposition rates and EC were recorded at Goma, highest values of Ca²⁺ and SO 4 2 − were recorded at Iko whereas the pH and PO 4 3 − were highest at Bukavu and high values of CO 3 2 − , NH 4 + , NO 3 − and K⁺ were recorded at Lwiro. The means of the ions were significantly different (P < 0.05) between the stations except for CO 3 2 − , where no significant difference was recorded (p > 0.05). The coefficient of variation was greater between all the stations except for pH which was 0.7 or less different.

The highest EC was recorded at Goma (47.4 ± 54 µS/cm) and the lowest at Lwiro (36.9 ± 52.8 µS/cm). The coefficient of variation varied between 1.1 and 1.4 in the stations. The highest pH was recorded at Bukavu (5.8 ± 0.4) and the lowest at Goma (5.6 ± 0.5). Total alkalinity calculated based on the [ HCO 3 − ] + 2[ CO 3 2 − ]− [H⁺] + [OH⁻] ratio, indicated that high alkalinity was recorded at Goma (126.9 mg/L) in dry atmospheric deposition and Bukavu in wet atmospheric deposition (258.5 mg/L). There is a statistically significant difference between Goma and Lwiro and Bukavu and Goma (p < 0.05). The mean pH showed that deposition was slightly acidic (pH < 6) at all four stations.

Wet atmospheric deposition results are presented in Table 3. The atmospheric deposition rate of major ions in wet atmospheric deposition varied slightly from station to station.

In the wet atmospheric deposition, the highest VWM concentration of Cl⁻,

CO 3 2 − , Ca²⁺, SO 4 2 − and PO 4 3 − were recorded at Lwiro. The mean was significantly different (P < 0.05) between the stations except for K⁺ (p > 0.05) and CO 3 − (p > 0.05). The coefficient of variation was greater between stations for ions VWM except for pH which was between 0.4 and 0.5.

High wet atmospheric deposition VWM concentration of the anion (Cl⁻, CO 3 2 − , SO 4 2 − and PO 4 3 − ) and cation (Ca²⁺) were recorded at Lwiro. High CO 3 2 − and Na⁺ was recorded at Bukavu, highest Mg²⁺, K⁺ and EC was recorded at Goma, while high NH 4 + and NO 3 − occurred at Iko and high HCO 3 − and Na⁺ were recorded at Bukavu. High EC was recorded at Goma 28.66 ± 13.0 µS/cm) and the lowest at Iko (17.4 ± 21.1 µS/cm). The pH was slightly acidic in all wet atmospheric deposition samples collected around the Lake Kivu basin (pH < 6). The total cation anion (TA/TC) ratios in wet atmospheric deposition revealed persistent free anions responsible for acidic precipitation in the region.

Enrichment factor (EF) was used to detect the source of major ions in dry and wet atmospheric deposition. The major ions ratios for dry and wet atmospheric deposition in different sites are given in Table 4.

Enrichment factors (EF) in the dry atmospheric deposition were generally low for the ratio of Na⁺ than in wet atmospheric deposition. However, high ratio was recorded except for K+, whose high values were recorded at Lwiro. Furthermore, EF factors of SO 4 2 − and Cl⁻ have high standard deviations in all the stations. Whereas the high EF factor of Ca²⁺ was different according to the ions taken,

relatively high EF factor values of Cl⁻ were recorded in samples at Goma, high Mg²⁺ occurred at Iko, high SO 4 2 − occurred at Lwiro, and high K⁺ was recorded at Bukavu.

For wet atmospheric deposition, EF calculated for the ions showed no significant difference (p > 0.05). Relatively high EF values of Na⁺ were recorded at Iko for Cl⁻, Mg²⁺, and SO 4 2 − ions and at Bukavu for Ca²⁺ and K⁺. While the EF ratios of Ca²⁺ were relatively high as recorded at Lwiro for Mg²⁺ and SO 4 2 − and at Goma for Cl⁻ and K⁺ ions.

A Pearson correlation analysis was performed to distinguish the possible common sources of ionic constituents. Table 5 shows the matrix of Pearson correlation coefficients (p < 0.05) among the major ions analysed in dry and wet atmospheric deposition

The ions in the dry atmospheric deposition that are strongly correlated include; CO 3 2 − , HCO 3 − , Mg²⁺, SO 4 2 − , PO 4 3 − and K⁺ while the others which are not correlated include NH 4 + , NO 3 − and Na⁺. These correlations are positive or negative between the parameters in dry atmospheric deposition. Correlations between ions in wet atmospheric deposition are presented in Table 6.

The ions in the dry atmospheric deposition that are strongly correlated include; HCO 3 − , Mg²⁺, NH 4 + , NO 3 − , SO 4 2 − , PO 4 3 − , Na⁺ and K⁺ while the others which are not correlated include CO 3 2 − and Ca²⁺. These correlations are positive or negative between the parameters in wet atmospheric deposition.

Although major ions in atmospheric deposition are essential for assessing sources of chemical influence, temporal trends in deposition are confounded by natural variability in wet and dry deposition. Temporal mean concentration (µeq/L) of some major ions in different sites of the Lake Kivu basin is depicted in Figure 2 for ions (Cl⁻, Ca²⁺, Mg²⁺, SO 4 2 − , PO 4 3 − , NO 3 − , Na⁺ and K⁺) in dry atmospheric deposition.

The high deposition concentration of major ions was recorded principally in January 2019 except for Ca²⁺ (December 2017) and NO 3 − (December 2018). But for all the major ions concentration, the trend remained similar in dry atmospheric deposition. Na⁺, Ca²⁺, K⁺ and SO 4 2 − were the dominant elements in the atmosphere around Lake Kivu basin in dry atmospheric deposition.

In the wet atmospheric deposition, the temporal mean concentration of some major ions in different sites in the Lake Kivu basin is presented in Figure 3.

High deposition concentrations of major ions were recorded principally in the rainy season. A high mean concentration of Cl⁻ was registered in January 2019, while high Ca²⁺ in June and SO 4 2 − and NO 3 − was recorded in April. A high rate of Mg²⁺ deposition was recorded in September. High concentration of Na⁺ (December), PO 4 3 − (April) and K⁺ (April) were recorded in these months respectively. This period corresponded with the period of rain in the area.

High mean concentrations of major ions in dry atmospheric deposition were recorded mainly in Goma for Cl⁻ and SO 4 2 − and Iko for Ca²⁺, Mg²⁺, PO 4 3 − , NO 3 − , Na⁺ and K⁺. This occurred in December to February and May to June for Goma. High rates of Cl⁻ were recorded in December and May, while those of; Ca²⁺ and SO 4 2 − were recorded in January 2018 and October, respectively. Those of PO 4 3 − and NO 3 − were recorded in January 2018 at Iko. Wet atmospheric deposition was high in the rainy season. High concentration of major ions such as (Cl⁻, Ca²⁺, Mg²⁺, PO 4 3 − , NO 3 − , Na⁺ and K⁺) were recorded at Lwiro, while high concentration for SO 4 2 − were recorded at Iko and for K⁺ at Bukavu. High concentration of major ions in the wet were recorded at Lwiro (Cl⁻, Ca²⁺, Mg²⁺, SO 4 2 − and NO 3 − ) and Iko ( PO 4 3 − and Na⁺).

Major ions that characterized atmospheric deposition were recorded in dry and wet atmospheric deposition. The relative dominance of NO 3 − , SO 4 2 − and Cl⁻ can potentially increase the acidity of the deposition although the dissolution of CO₂ in the samples would have already reacted with distilled water. The ratios of SO 4 2 − + 2NO 3 − /Ca²⁺ + Mg²⁺ showed alkaline water recorded in dry and wet atmospheric deposition respectively. The presence of Ca²⁺, NH 4 + , HCO 3 − and CO 3 2 − contributed to the neutralization process of acid in atmospheric deposition [22]. Whereas the presence of NO 3 − , SO 4 2 − and Cl⁻ increased the acidity in atmospheric deposition. The results revealed that wet and dry atmospheric deposition samples were acidic at all the four stations with mean pH values varied

between 5.19 and 5.57 in wet atmospheric deposition, and 5.56 and 5.82 in dry atmospheric deposition. I dry atmospheric deposition alkalinity varied from location as follows; Goma (126.7 ± 3.1 mg/L) > Iko (105.9 ± 2.1 mg/L) > Bukavu (105.8 ± 4.5 mg/L) > Bukavu (99.6 ± 3.1 mg/L) respectively; while in wet atmospheric deposition, it decreased as follows; Bukavu (258.5 ± 4.5 mg/L) > Goma (199.1 ± 3.1 mg/L) > Iko (152.2 ± 2.1 mg/L) > Lwiro (90.3 ± 3.1 mg/L) respectively.

Low mean pH wet deposition values were recorded at Goma station (5.19) compared to other stations. Bicarbonate ion rate is important for these values of pH obtained in the samples. This was a result of the acid-base balance of all chemical species in the rain. In a poorly buffered atmosphere, acid deposition leads to acidification and has a negative impact on the ecosystem as found by Wang et al., [23] and Akimoto et al., [24] in China and Japan respectively.

Except for Ca²⁺ and SO 4 2 − dissolved major ions were close to those reported for other areas worldwide [25]. The most abundant anions in the dry and wet atmospheric deposition were SO 4 2 − , similar to the finding by Derry and Chadwick [26]. The most abundant cation was Na⁺ followed by K⁺, contrary to the results observed by Bootsma et al. [27] in the Lake Malawi basin where Ca²⁺ was the abundant cation.

Stefánsson et al. [28] found that major ions were present for regions with active volcanoes which compared well with the results of this study. Volcanoes are considered the most important natural sources of emissions of major ions. In contrast, the most important anthropogenic sources are fossil fuel combustion for energy production, transportation, factories activities in big cities and their surroundings in the basin. However, biomass burning from deforestation, land clearing and bush fires are also potential sources of major ions in the atmosphere [29] [30]. In the Lake Kivu basin, the presence of active volcanoes of Virunga contributes to SO 4 2 − and NO 3 − in dry and wet atmospheric deposition. Moreover, the presence of the cement factory of Katana (CIMENKI/Katana) and different lime production enterprises located near the lake are also potential sources of atmospheric SO 4 2 − , HCO 3 − and CO 3 2 − in the atmosphere of the Lake Kivu basin, which increase the concentrations in the atmosphere.

The factory releases dust in the atmosphere containing different chemical constituents such as SO 4 2 − with a longer lifetime that, when deposited, can lead to being spread over greater distances [31]. This high atmospheric deposition of SO 4 2 − can be explained in the region by the mineralogy of the soil material. The presence of calcite (CaCO₃), dolomite (CaCO₃∙MgCO₃) and gypsum (CaSO₄∙2H₂O) in the soil confirm the high atmospheric deposition of major ions in the atmosphere. The dissolution of these minerals’ accounts for the high VWM of Ca²⁺ and SO 4 2 − . These minerals were found in dust material in other regions and are used in the cement factory in CIMENKI/Kakondo and lime production enterprises in the basin. The high atmospheric deposition VWM of SO 4 2 − and moderate Ca²⁺ at Iko station was probably due to the influence of the cement factory located near the site at Kakondo.

The atmospheric deposition rate of these ions (Ca²⁺, SO 4 2 − and Mg²⁺) are associated with wind direction and transport mechanisms in the basin. During rain events, generally high atmospheric deposition rates of Ca²⁺ and SO 4 2 − were recorded in the Lake Kivu basin similar to the observation of other authors in different regions [2]. The high rate may be due to the process of scavenging during the precipitation. For the station of Goma, major sources of ions are expected to be volcanic ash in the atmosphere but also an uncompleted road network where dust is usually blown when traffic is high. In the dry atmospheric deposition, high rates of major ions of SO 4 2 − and CO 3 2 − were recorded at the station of Goma. The occurrence of relatively high levels of anthropogenic activities and the volcanoes are the factors of high species of NO 3 − , SO 4 2 − observed in this station. This was also observed in active volcanic areas like Indonesia, Costa Rica, and Nicaragua [32].

However, studies in central Africa and South America (e.g., [8] [33] [34] ), found that the principal source of atmospheric deposition was the emission from burning forests, savannas and agricultural fields as observed in this study. As Wang et al. [25] reported, it is worthwhile to note that the proportion of Ca²⁺ in mineral dust varied from one site to another. Bootsma and Hecky [35] found that soil dust resulting from deforestation and agriculture was responsible for increased P deposition in Uganda.

In general, temporal variability was recorded in the VWM of major ions for dry and wet atmospheric deposition in the four stations. High atmospheric deposition VWM was recorded in the rain season. Such temporal variation was also observed at Mt Etna in Italy [36]. At Bukavu, the dust blown from road traffic spreads out in the area, contributing to the ions into the atmosphere returning as dry and wet deposition. Other sources of pollutants in the atmosphere are anthropogenic activities such as burning waste and emissions from waste disposal sites. The same trend was observed at Goma station because the two stations are urbanized and characterized by high population and high traffic on the road. In addition, the Goma station is located in an active volcanic area. Emission of volcanoes in this region increases major ions in dry and wet atmospheric deposition. However, volcanic ash known as an aerosol contains major ions such as SO 4 2 − , NO 3 − , Na⁺, K⁺, Ca²⁺ and Mg²⁺ [36]. Iko station is located on the island near the cement industry in the coastal area. Generally, the rain was unevenly distributed in the basin during sampling, and some stations received less water. The variability followed the trend of the activities and the seasonality in the basin, as observed in various China regions [37].

The strong positive correlation coefficient in dry atmospheric deposition between both SO 4 2 − and Ca²⁺ with some ions (Table 5) can be explained by the presence of cement industry and other anthropogenic activities. These activities in the region included road construction, blowing soil in the atmosphere, lime fabrication enterprise, forest cutting and agricultural practice in the Lake Kivu basin. In this basin, the presence of cement industry and lime fabrication enterprises near the site at Iko was the main source of SO 4 2 − and Ca²⁺ with high deposition rates in dry deposition at Iko and Lwiro. Nitrate was positively correlated with Na⁺ and K⁺ with possible contribution arising from biomass burning and blowing aerosol which are common activities in the basin. Meanwhile, Bukavu and Lwiro, have high deposition rates of NO 3 − than in other sites. In India, Khare et al. [38] found that K⁺ ion was associated with airborne particles generated by biomass burning in the basin that was also observed in this study. There was a correlation with other ions that are produced by burning in dry deposition, such as NO 3 − .

According to Coelho et al. [39], vehicle emissions substantially increase the atmospheric deposition of major ions, as observed in samples collected in Turkey compares with our findings in the Lake Kivu basin. In addition, the influence of local road traffic, which enhances the suspension of soil-derived particles into the atmosphere, have contributed to such increase of particles as found by Safai et al. [20]. It is important to note that dust carried by wind and soil plays significant role in dry deposition chemistry. The presence of several limestone mines and the cement industry in the basin contribute to the increased rates of some specific ions in dry atmospheric deposition.

The Mg²⁺, Ca²⁺, SO 4 2 − ions in wet atmospheric deposition suggested the presence of crustal contributions or from cement manufacturing. This behaviour of Mg²⁺ was also found in rainwater in China [37]. The dissolution of the carbonate minerals such as calcite and dolomite, a mineral that is composed in the soil in the region, was the essential reaction that affected the rainwater composition, as found by Baker and Jickells [8]. SO 4 2 − ions are well correlated with NH 4 + and PO 4 3 − and negatively correlated with pH and NO 3 − was correlated with Na⁺, K⁺ and EC but also negatively correlated with pH, suggesting that the acidic pH originated from the anthropogenic input. pH was only correlated to CO 3 2 − in wet atmospheric deposition, which supported the contribution of acidity coming from H₂SO₄ and HNO₃, which was lower than the free acidity observed due to the neutralization process was responsible of acidification [40].

Sulphate and NO 3 − ions were not correlated, suggesting that the contributions of the observed deposition of these ions were either from volcanic ash or soil blowing in the atmosphere. The processes such as wind erosion, burning and tillage caused particles containing major ions to become airborne [33]. These findings confirm the influence of natural and anthropogenic sources of these ions in wet deposition around the Lake Kivu watershed. The ammonium ion is notable with its poor correlation with other major ions. It is negatively correlated with pH, suggesting that some NH 4 + are not associated with volcanic sources but reflect agricultural activity’s influence. Sodium and Mg²⁺ ions were strongly correlated, suggesting an influence from soil dust. Coelho et al. [39] indicated that NaCl, CaSO₄, MgSO₄, MgCl₂, HNO₃, NH₄NO₃, NH₄HSO₄ and (NH₄)₂SO₄ were the predominant species combinations that could occur in the atmosphere by following the major ions in contact in the atmosphere. They could have been formed by scavenging of aerosols and reaction of gaseous species in water droplets.

Enrichment factors (EF) were a useful tool in understanding atmospheric deposition in Lake Kivu. The composition of the bedrock revealed that calcium was the dominant natural element [41]. Enrichment factors of the different ions with Ca²⁺ showed a similar trend with the corresponding soil ratio EF between 0.73 - 1.94. According to Ma et al. [42], most of the EFs of other elements were below or close to ten, suggesting an influence from soil particles.

The EF analysis confirmed that SO 4 2 − , Cl⁻, Ca²⁺ and Mg²⁺ originated from soil and volcanic ash, biomass burning, cement industry and lime production enterprises. At the same time, Na⁺ came mostly from biomass burning, and the more significant part of K⁺ was from the lake (marine origin). The application of factor analysis reveals that crusting and anthropogenic activities played an important role in the distribution of ionic species in wet precipitation in the Lake Kivu basin.

Generally, lake contribution of NO 3 − and NH 4 + was minimal; therefore, EFs for the lake source were computed for Cl⁻, Mg²⁺, K⁺, SO 4 2 − and Ca²⁺, while EFs for soil source were computed for K⁺, Na⁺, Mg²⁺, Cl⁻, SO 4 2 − and NO 3 − . The EF approaching unity suggests that the elements originate from the soil. If the EF was greater than unity, it implies that a significant fraction had an origin in a different than soil/lake. Main sources of Ca²⁺, Na⁺ and Mg²⁺ are known to be obtained during the crusting processes and wind-blown soil [17] as revealed in Table 4. The EFs of these elements were higher than the corresponding EF in the soil in the region. These were also confirmed in the study of Barbieri et al. [43], who found the highest deposition concentration of Ca²⁺ ion in soil near the sampling site. The potential contribution of the cement industry and other small enterprises producing lime in the region were also contributors of Ca²⁺ ions in the dust to the atmosphere by local wind systems.

According to Rastegari et al. [44], the ratios Na⁺/Ca²⁺, SO 4 2 − /Ca²⁺ and K⁺/Ca²⁺ EF indicated a dominant soil origin for both species, while K⁺, Mg²⁺, SO 4 2 − , Na⁺ and Ca²⁺ contents pointed to additional sources (soil dust, fertilizers, etc.) as nitrate and sulphate displayed a nearly constant EFs indicating a common anthropogenic origin. This was also found in Lake Kivu basin where activities such as agricultural practices and cement industry are common leading to susceptibility of increased biomass burning and dust in the atmosphere, concurring with observations by Naimabadi et al. [45].