Metal and semimetal loadings in sediments and water from mangrove ecosystems: A preliminary assessment of anthropogenic enrichment in São Tomé island (central Africa)

Chemosphere 334 (2023) 138973 Available online 19 May 2023 0045-6535/© 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Metal and semimetal loadings in sediments and water from mangrove ecosystems: A preliminary assessment of anthropogenic enrichment in S˜ ao Tom´ e island (central Africa) F. Afonso a , * , C. Palma b , ** , A.C. Brito a , c , P. Chainho a , d , R. de Lima e , f , g , h , J.A. Heumüller a , F. Ribeiro a , P.M. F´ elix a a MARE – Marine and Environmental Sciences Centre/ ARNET - Aquatic Research Network, Faculdade de Ciˆ encias, Universidade de Lisboa, Campo Grande, 1749-016 Lisboa, Portugal b Instituto Hidrogr´ afico, Rua das Trinas 49, 1249-093 Lisboa, Portugal c Departamento de Biologia Vegetal, Faculdade de Ciˆ encias da Universidade de Lisboa, Campo Grande 016, 1749-016 Lisboa, Portugal d CINEA and ESTS, IPS – Energy and Environment Research Center, Instituto Polit´ ecnico de Setúbal, Estefanilha, 2910-761 Setúbal, Portugal e Departamento de Biologia Animal, Faculdade de Ciˆ encias da Universidade de Lisboa, Campo Grande 016, 1749-016 Lisboa, Portugal f cE3c - Centre for Ecology, Evolution and Environmental Changes, Faculdade de Ciˆ encias da Universidade de Lisboa, Campo Grande 016, 1749-016 Lisboa, Portugal g CHANGE - Global Change and Sustainability Institute, Faculdade de Ciˆ encias da Universidade de Lisboa, Campo Grande 016, 1749-016 Lisboa, Portugal h CBGG - Centro de Biodiversidade do Golfo da Guin´ e, S˜ ao Tom´ e, S˜ ao Tom´ e Island, Sao Tome and Principe HIGHLIGHTS GRAPHICAL ABSTRACT •High concentrations of Cu and Ni in S˜ ao Tom´ e mangroves have a lithogenic origin. •Contamination by As and Cr suggest a strong contribution from human activities. •Smallest systems had higher metal content in water, but no sediment contamination. •Higher water contamination represents the continuous input from waste and chemicals. •Non-industrialized landscapes may also show severe contamination by human activities. ARTICLE INFO Handling Editor: Giulia GUERRIERO Keywords: Geochemical index Health risk Metal contamination Pollution source ABSTRACT Mangroves act as buffer areas for marine systems, providing a barrier to contamination from continental sources by retaining metal pollutants. This study evaluates metal and semimetal contamination in the water column and sediments of four mangroves located on the volcanic island of S˜ ao Tom´ e. Several metals had a widespread distribution, with occasional high concentrations, linked to potential sources of contamination. However, the two smaller mangroves, located in the northern part of the island, tended to have high metal concentrations. Arsenic and chromium concentrations were notably concerning, particularly if we consider this is an isolated and non- * Corresponding author.MARE – Marine and Environmental Sciences Centre, Faculdade de Ciˆ encias da Universidade de Lisboa, Campo Grande, 1749-016 Lisboa, Portugal. ** Corresponding author. E-mail addresses: [email protected] (F. Afonso), [email protected] (C. Palma). Contents lists available at ScienceDirect Chemosphere journal homepage: www.elsevier.com/locate/chemosphere https://doi.org/10.1016/j.chemosphere.2023.138973 Received 1 March 2023; Received in revised form 11 May 2023; Accepted 16 May 2023 Chemosphere 334 (2023) 138973 2 Transitional aquatic systems Volcanic enrichment industrialized island. This work highlights the need for further assessments and a better understanding of processes and implications of metal contamination in mangroves. This assumes a particular relevance in areas that have specific geochemical compositions (i.e., volcanic origin) and in developing countries, where people often rely directly and heavily on resources obtained from these ecosystems. 1. Introduction Mangroves are transitional aquatic systems that connect terrestrial, riverine and marine systems. They are mostly found in developing countries located in tropical areas, and provide a considerable number of services, not only locally (Afonso et al., 2021) but also on wider scales (Bayen, 2012). One of the valuable services they provide is protection from metal pollution and other contaminants due to their buffering capacity. As receivers of continental run-off, they can influence the occurrence and availability of metals, due to their ability to trap these elements, thus, halting the contamination of adjacent marine ecosystems (Harbison, 1986; Schaffelke et al., 2005). Mangrove sediment is rich in sulphide and organic matter (Silva et al., 1990; Tam and Wong, 2000), which favours the retention of metals that form complexes with sulphides (Clark et al., 1998). Despite the ability of plant roots to regulate metal uptake, these complexes tend to facilitate metal mobilization and bioavailability for root uptake, favouring accumulation in the plants (Clark et al., 1998; MacFarlane and Burchett, 2002). While sediments tend to trap metals, the water column transports them, but remobilization can convey metals from the sediment back to the water column in a dynamic cycle (Caccia et al., 2003; MacFarlane et al., 2007; Zhang et al., 2017). This means that remobilization can transform the sediment into a secondary source of metal contamination of the water column (Bodin et al., 2013). Nevertheless, there is a gap in knowledge of tropical ecotoxicology, namely regarding the response of mangroves to the intrusion of metals (Duarte et al., 2019). Metal occurrence in ecosystems can have natural or anthropogenic sources, and this distinction is important when assessing environmental risk (Dias and Edwards, 2003). Apart from soil and rock weathering, volcanism is also an important natural source of metals (Gnandi et al., 2011). Metal pollution is considered to be one of the most damaging types of anthropogenic contamination, due to their toxicity, persistence, ubiquity, non-biodegradability, and bioaccumulation (Duman et al., 2007; MacFarlane et al., 2007). In mangroves, anthropogenic metal contamination has been linked to urbanization, agriculture run-off, industrial effluents, leaching from domestic garbage dumps, boat emissions, and anti-fouling chemicals from boat paint (Peters et al., 1997). Counterintuitively, wetland contamination is not necessarily more severe in developed countries. The Niger Delta, Nigeria, showcases this proposition, as it is considered the most severely damaged ecosystem in the world (Kadafa, 2012). Some metals are essential to human health, while others can be toxic even at low concentrations. When introduced in ecosystems on which people are strongly dependent, these elements can have serious effects on the physiological processes of the human body, leading to severe health consequences (Bosch et al., 2016; Rehman et al., 2018). Apart from ingestion, metals can also enter the human body by inhalation or dermal contact (F´ elix, 2012). In children up to six years old, one specific gateway for metal intake is the hand-to-mouth behaviour that leads them to ingest 39–271 mg day −1 of soil (Ljung et al., 2007; Wang et al., 2021). Considering the potential damage that metals can have on human health and ecosystems, it is essential to improve environmental risk assessment. There are several tools to assess metal contamination, and to distinguish lithogenic and anthropogenic sources. Several empirical methodologies are used to assess the status of metal pollution, and to estimate the impact of human activities on ecosystem health (Alzahrani et al., 2018; Chabukdhara and Nema, 2012). Of these, the geochemical indices, such as the contamination factor and the enrichment factor, compare field measurements to standard values to determine a contamination degree (Brady et al., 2015; Nath et al., 2014). The enrichment factor compares metal concentrations to those of a reference situation, allowing the quantification of human-induced changes in the ecosystem (OSPAR Commision, 2009). The dominant factor for determining the metal content in soils is the parent rock, and volcanic soils tend to show high concentrations of metal elements (Brtnický et al., 2020), namely chromium, mercury, and cadmium (Ahn and Chon, 2010; McMurtry et al., 1995). This study aims to determine metal and semimetal concentrations and potential anthropogenic enrichment in the sediment and in the water column of mangroves located on the volcanic island of S˜ ao Tom´ e, Gulf of Guinea. Four mangroves with contrasting features are used to compare anthropogenic contamination. This information is vital to start identifying hazardous effects of potential high concentrations of metals on people and ecosystems, and to define strategies to maximize the capacity of mangroves to sink metallic elements. 2. Materials and methods 2.1. Study area The study was developed on the volcanic island of S˜ ao Tom´ e (S˜ ao Tom´ e and Príncipe, Gulf of Guinea, central Africa), which is predominantly composed of recent basaltic lavas (Caldeira and Munh´ a, 2002). Four mangroves (Fig. 1), displaying contrasting environmental contexts and levels of anthropogenic pressure, were selected: Malanza, Angolares, Praia das Conchas, and Diogo Nunes (Table 1). Malanza (Fig. 1) and Praia das Conchas (Fig. 1) are both in distinct exclaves of the Obˆ o Natural Park, and both are open systems that have constrictions in their connections to the sea (F´ elix et al., 2017). Diogo Nunes (Fig. 1), located very close to the airport, is not in a protected area, it has the largest hydrographic basin of all studied mangroves and it is an intertidal mangrove that is dominated by saltwater intrusions during the high tide, and freshwater on low tide. Angolares (Fig. 1) is also not in a protected area. It conveys two separate contiguous hydrographic basins and, thus, has two tributaries draining into the mangrove, creating two arms. The southern arm is permanently open, while the northern one is a lagoon that is periodically closed by a sand barrier due to the accumulation of sediments by tidal influence. This barrier breaks periodically, mostly during periods of strong rains, due to riverine influence. Water accumulation builds pressure and, eventually, breaks the sand barrier, which can remain open for a variable period of time, depending on the balance between freshwater and coastal dynamics, as in enclosed lagoons (F´ elix et al., 2013). The northern arm of this system is closest to the S˜ ao Jo˜ ao dos Angolares city, easily accessible to locals for fishing, bathing, and litter dumping (Afonso et al., 2022). At S˜ ao Tom´ e, rivers are often commonly used as laundry areas, where soaps and detergents are used every day (Afonso et al., 2022). Most products from this activity enter the mangroves via run-off, but at Praia das Conchas and Diogo Nunes laundry takes place directly in the mangroves, whenever these are dominated by freshwater. Both these systems are also highly used for trash dump. The I´ ogoi´ ogo bay is a marine area adjacent to the Malanza mangrove, which is intensively used by the nearby communities, notably to support their fishing activities, and has a strong polluting potential, namely considering that waste products from these activities are commonly dumped at the port area. Two main docks were identified in this bay: the F. Afonso et al. Chemosphere 334 (2023) 138973 3 fishing boat harbour at Porto Alegre , where boat and gear repair occur, but also washing, painting and garbage dumping; and the Ponta Baleia docking area, where most passengers and goods are transported to and from the Ilh´ eu das Rolas resort, 2.5 km south of the main island of S˜ ao Tom´ e (Fig. 1). 2.2. Sample collection and preparation Water and sediment sampling occurred in August 2017, representing the systems’ environmental variability and especially the estuarine longitudinal gradient. Sampling effort depended mostly on mangrove size and variability, with sampling stations selected at Praia das Conchas, Diogo Nunes Angolares, Malanza, and at I´ ogoi´ ogo bay (Fig. 1). All reagents used were ultra-pure (u.p.), and u.p. water (MilliQ, 18.2 MΩ.cm purified by a Q-POD® Element Merck Millipore) was used in the preparation of all solutions. All sampling containers were previously washed using 10% nitric acid and then rinsed with ultra-pure MilliQ water. A total of 32 water samples were collected at all sampling stations: 3 samples from Praia das Conchas, 3 samples from Diogo Nunes, 6 samples from Angolares, 7 samples from Malanza and 13 samples from I´ ogoi´ ogo bay. The samples were collected directly to 1 L polyethylene containers, and then acidified with concentrated HNO 3 to a pH <2, and refrigerated at 4 ◦C until analysis. Before collection, containers were rinsed three times with water from the sampling location, immediately before collection. Samples were filtered through pre-cleaned Whatman WCA 0.45 μ m filters. A total of 19 surface sediment samples were collected: 3 samples from Praia das Conchas, 4 samples from Diogo Nunes, 5 samples from Angolares and 7 samples from Malanza. The top 3 cm of the sediment were collected directly into 2 different polyethylene containers, by pressing the container, upside-down, against the sediment, carefully scooping the material and covering the containers without manipulation. All samples were frozen between collection and analysis. Total organic matter (TOM) in the sediment was obtained by loss on ignition (EN15169, 2007). One of the containers was used to determine the granulometry by the separation of dried samples with calibrated sieves and classified into silt (<0.063 μ m), fine sand (0.063 μ m ≤x <0.250 μ m), medium sand (0.250 μ m ≤x <0.500 μ m) and coarse sand (0.500 μ m ≤x <2 mm), according to (Blott and Pye, 2001). 2.3. Metal analysis To analyse metal concentration in the water (Cd, Cu, Fe, Mn, Ni, Pb and Zn), samples were subjected to a pre-concentration by solid-phase extraction (SPE) using selective cartridges that retain the metals while general cations of the saline samples are eliminated. To optimize the extraction, samples were previously adjusted to a pH of 6.5, using Fig. 1. S˜ ao Tom´ e Island (main map on the left), showing the location of the mangroves targeted for this study (Praia das Conchas, Diogo Nunes, Angolares, and Malanza), detailed on the right-hand insets. The inset on the top left shows the location of S˜ ao Tom´ e in Africa. F. Afonso et al. Chemosphere 334 (2023) 138973 4 ammonia acetate (0.5 M), ammonia (20–22%) and nitric acid (2 M). The cartridges were placed in the vacuum extraction system (Manifold), conditioned with HNO 3 (2 M), u.p H 2 O and ammonia acetate (0.5 M). Samples were then placed in the cartridges, which were washed with H 2 O u.p.. The elution of metals with HNO 3 (2 M) was then performed. This procedure follows that of the DigiSep Blue Label cartridges from SCP Science (ref: CATALOG NUMBER MK-MKG003-SPEB-2.0-E). Quantification was done by Atomic Absorption Spectroscopy (AAS) in a Solaar–Thermo Elemental equipment. Cd, Pb and Ni were analysed by graphite chamber, and Cu, Fe and Zn by flame. The calibration curve was obtained from external standards, and the working solutions of metals were prepared using 1000 mg L −1 (Merck) standard solutions. The detection limit was 0.01 mg L −1 for Ni, Cd and Pb, 0.1 mg L −1 for Cu, Mn and Zn, and 0.3 mg L −1 for Fe. The quality control procedure included analysis of duplicate samples (one for every batch of ten samples), blanks between samples, recovery tests and participation in the intercalibration exercise QUASIMEME (Quality Assurance of Information in Marine Environmental Monitoring) for external quality control. The results obtained during the period of this analysis showed a zscore <1, the average recoveries of the standard solutions were between 80 and 110% for all the metals, which confirmed the accuracy of the method. The precision of the method was confirmed by the duplicate samples. To assess metal concentration in the sediment (same elements analysed in the water plus Al and Cr), the <2 mm fraction (using a calibrated sieve) was freeze-dried, homogenised, and reduced to a fine powder using a mortar and pestle (according to OSPAR Commission, 2015). Sediment aliquots of 0.45–0.55 g were placed in Teflon vessels (2 ml HF (40%) and 6 ml of aqua regia – HNO 3 (65%) and HCl (30%) (1:3), and then acid-digested in a Milestone® ETHOS PLUS microwave oven . After digestion, each solution was neutralised with boric acid and diluted to 50 ml with ultrapure water. Metal concentrations were analysed via flame atomic absorption spectrometry in a Thermo Scientific iCE™ 3500 (EPA 3051A standard, 2007; OSPAR Commission, 2015), except for As, which was analysed in a hydride generator, with sodium borohydride added to convert As (III) to the volatile hydride that was then purged from the solution by a stream of argon gas in a Thermo Elemental-VP90 Continuous Flow vapour Accessory (ISO 17378-2:2014 E). The calibration curve was obtained by external standards, and the working solutions of metals were prepared using 1000 mg L −1 (Merck) standard solutions. The detection limits were 1.6 mg kg −1 for Cu, Fe, Mn and Cr, 0.6 mg kg −1 for Zn, 16 mg kg −1 for Al, 3 mg kg −1 for Pb, 2,5 mg kg −1 for Ni, 0,16 mg kg −1 for Cd and As, and 2.7 mg kg −1 for Hg. All reagents were of Merck Suprapure quality and the u.p water used . The analytical procedure included one reagent blank in each batch of 20 samples, analysis of duplicate samples (one for every batch of ten samples), recovery of standard solutions, certified reference materials (CRMs), and participation in the QUASIMEME for external quality control. CRMs (LGC 6187 and MESS-4) were analysed following the same procedure. Certified and measured values matched, with recovery between 80% and 120% for all CRMs and metals. Participation in 50 QUASIMEME exercises (further demonstrating data quality, with more than 90% acceptable results (z-score <2)). This analysis confirmed the precision and accuracy of the methods for all the metals. All metal concentrations were calculated on a dry weight basis, and water content was measured by drying known amounts of sediment at 105 ◦C. 2.4. Data analysis Average sediment metal concentrations obtained in S˜ ao Tom´ e mangroves was compared to those from other mangroves and volcanic islands, compiled from the literature. The potential for anthropogenic input of metals into the environment was assessed through several widely used methods, only available for sediment concentrations: The Enrichment Factor (EF), the Geoaccumulation Index (Igeo), and the Contamination Factor (CF). The EF Table 1 Features of the four mangroves studied in S˜ ao Tom´ e Island (adapted from Afonso et al., 2021; Caldeira et al., 2003). Dash indicates that this value was not available. Mangrove Malanza Angolares Diogo Nunes Praia das Conchas Mangrove area (km 2 ) 2.83 0.68 0.12 0.0012 Area of the hydrographic basin (km 2 ) 7.41 7.82 23.35 13.31 Geologic composition of the hydrographic basin(s) Basalt-trachyphonolite lavas; alluvial deposits; pyroclastic and lava cones Basalt-trachyphonolite lavas Basaltic lavas; pyroclastic and lava cones Basaltic lavas; alluvial deposits Size of the closest human settlement (habitants) 1345 2037 392 180 Land use dominance in the surroundings Mangrove (53.6%) Agroforestry (59.4%) Agriculture (47.5%) – F. Afonso et al. Chemosphere 334 (2023) 138973 5 differentiates anthropogenic and natural sources of sediment enrichment for each element (Szefer et al., 1999). It was calculated using a formula that uses Al as the normalizer of metal concentrations: EF =(C/CAl)sample (C/CAl)reference where (C/C Al ) is the ratio of concentration of the element of concern (C) to that of Al (C Al ), in both the sediment sample (sample) and in an unpolluted reference sample (reference). In the present study, the reference sample values were taken from the bibliographic reference of shallow water sediment that was closest to the conditions found in the study area (Salomons and Forstner, 1984). Five contamination categories are acknowledged on the basis of the enrichment factor: 2 <EF <5 - deficiency to moderate enrichment; 5 <EF <10 - moderately severe enrichment, 10 <EF <25 - severe enrichment; 25 <EF <50 - very severe enrichment; and EF >50 - extremely high enrichment (Birch and Olmos, 2008). The Igeo quantifies the degree of metal pollution (Müller, 1979), and has been used in studies of trace metal studies in sediments and soils (e. g., Amin et al., 2009a). It was calculated as: Igeo =log2(Cn 1.5Bn) where C n is the concentration of a metal in the sediment, and B n is the geochemical background value of that same metal in shallow water sediment (Müller, 1979). The Igeo is then classified into seven classes: 0 (practically uncontaminated) - Igeo<0; 1 (uncontaminated to moderately contaminated) - 0 <Igeo<1; 2 (moderately contaminated) - 1 <Igeo<2; 3 (moderately to heavily contaminated) - 2 <Igeo<3; 4 Fig. 2. Metal concentrations ( μ g/L) in the water of S˜ ao Tom´ e studied mangrove systems. Sites with unusually high concentration are tagged. F. Afonso et al. Chemosphere 334 (2023) 138973 6 (heavily contaminated) - 3 <Igeo<4; 5 (heavily to extremely contaminated) - 4 <Igeo<5; 6 (extremely contaminated) - 5 <Igeo. The CF expresses the level of metal contamination of sediment, being calculated as: CF =Cn Bn where C n is the concentration of a given metal in the sample, and B n is the geochemical background value of that same metal in shallow water sediment (Müller, 1979). CF values were interpreted as: low contamination for CF <1; moderate contamination for 1 <CF <3; considerable contamination for 3 <CF <6; and very high contamination for CF >6 (Hakanson, 1980). 3. Results 3.1. Metal concentration in the water Metal concentrations in the water column were highly variable between sites and elements, with several sites highlighted by high concentrations for some elements (Table S1 – Fig. 2). The highest concentrations in the water were the following: upstream in Malanza for cadmium (M6 – 0.73 μ g/L); downstream in Diogo Nunes (D1 – 3.46 μ g/ L) and in Praia das Conchas (PC1 – 2.99 μ g/L), and near the fishing harbour in I´ ogoi´ ogo Bay for copper (B0 – 3.47 μ g/L); upstream in Malanza for iron (M7 – 4345.68 μ g/L) downstream in Diogo Nunes (DN1 – 134.44 μ g/L) and Praia das Conchas (PC1 – 139.08 μ g/L), and upstream in Malanza for manganese (M6 – 142.03 μ g/L, M7 – 121.60 μ g/ L); downstream in Diogo Nunes (D1 – 46.79 μ g/L) and in a small shallow pound in Malanza for nickel (M1 – 13.57 μ g/L). The highest lead concentrations were found in one of the arms of Diogo Nunes, and near the passenger docking area and north section of Malanza bay (DN3 – 2.45 μ g/L, B0 2.64 μ g/L, B5 3.58 μ g/L, Tables S1 – Fig. 2). Zinc concentrations were higher in water samples from Angolares north arm, upstream area (A4 – 19.33 μ g/L, Table S1 – Fig. 2), in the small shallow pond of Malanza (M1 – 15.51 μ g/L, Table S1 – Fig. 2), and near the fishing harbour in Malanza bay (B0 – 13.83 μ g/L, Table S1 – Fig. 2). The comparison to background values of metal content (nickel, cadmium, lead, copper, and zinc) showed that S˜ ao Tom´ e mangrove water had higher concentrations of copper (Table S1). Praia das Conchas and Diogo Nunes presented concentrations of nickel, cadmium and zinc that were particularly higher than the reference values (Table S1). 3.2. Metal loadings in surface sediments 3.2.1. Sediment characterization: granulometry and organic matter content Granulometry of surface sediment was highly variable between and sometimes within mangroves, even though it was mostly composed by sands and silts (Fig. 3). All samples from Diogo Nunes were classified as sand with small amounts of organic matter. The downstream areas of Angolares from both arms were predominantly sandy (99% at A1, and 98.8% at A2), while the northern arm had muddy sand in the middle (A3, A4 – 54% and 67% sand respectively) and sandy gravel upstream (A5 – 20% sand). In Malanza, the sediment was very homogenous in upstream area, being mostly muddy sand (61% sand at M5, 56% at M6, and 65% at M7) with high amounts of organic matter (42%, 61% and 63% respectively), while downstream locations (99% sand at M2, 85% and M3, and 85% at M4) were mostly sandy with a small proportion of organic matter (7%, 11% and 13%, respectively). Praia das Conchas had muddy sand upstream (80% sand at P3), sandy gravel in the middle (51% gravel and 48% sand at P2), and sand downstream (98% sand at P1). Fig. 3. Granulometry ternary diagram. Each colour represents a sampling location, and the size of the circles indicates organic matter content (i.e., larger circles indicate higher amounts of organic matter). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) F. Afonso et al. Chemosphere 334 (2023) 138973 7 3.2.2. Comparing sediment metal concentration in S˜ ao Tom´ e mangroves with similar systems We started by comparing average elemental concentrations obtained in S˜ ao Tom´ e mangroves with those from other mangrove and volcanic islands, compiled from the literature. Since values with aluminium normalization are mostly applied in sediments with marine origin (Herut and Sandler, 2006), the comparison was based only on absolute concentrations. S˜ ao Tom´ e mangrove sediments tended to present higher mean metal concentrations when compared to other mangroves and volcanic islands, except for lead in other mangrove systems and copper and nickel in other volcanic islands (Table 2). 3.2.3. Enrichment and contamination Enrichment (EF) and contamination (Igeo and CF) based on metal concentrations in the surface sediment (Table S2), were highly variable between sites and elements, also showing relevant differences between indices (Table 3). Cadmium concentrations in the sediment were very close to the quantification level, therefore the indices were not calculated. S˜ ao Tom´ e mangroves had consistently low enrichment and contamination by copper, lithium and lead, while iron, mercury and zinc had only occasional high values of enrichment, upstream in Malanza (M6). Manganese enrichement was found at all sampling points of Diogo Nunes, upstream at Malanza (again, M6) and downstream at Praia das Conchas (P1), while nickel enrichement was found upstream at Diogo Nunes (D3 and D4) and across most sampling points of Malanza (except M6 and M7) and showed occasional contamination at the small contiguous pond in Malanza (M1) and upstream at Praia das Conchas (P3). Only arsenic and chromium presented a widespread enrichment and contamination, and very severe enrichment and heavy contamination at a few locations. Arsenic enrichment and contamination were particularly noticeable at Diogo Nunes and Malanza. 4. Discussion The assessment of the metal loadings in four mangroves in S˜ ao Tom´ e Island depicted, generally, the same patterns as in those found in West African regions, which are generally richer in cadmium, chromium, nickel, and zinc, especially in coastal areas (Gnandi et al., 2011). Praia das Conchas and Diogo Nunes even being the smallest study areas were characteristically the systems that presented higher metal content. On the other hand, Angolares presented the lowest enrichment and contamination levels. Table 2 Mean sediment metal concentrations (absolute values, mg kg −1 ) for S˜ ao Tom´ e mangroves (current study), and in other mangrove systems and volcanic islands (taken from the literature). Values highlighted in grey show mean concentrations that were lower than those found in S˜ ao Tom´ e mangroves. Dashes indicate that those values were not available. [1] (Kruitwagen et al., 2008); [2] (Gnandi et al., 2011); [3] (Cuong et al., 2005); [4] (Essien et al., 2009); [5] (Amin et al., 2009b); [6] (Vane et al., 2009); [7] (Marchand et al., 2006); [8] (Preda and Cox, 2002); [9] (Coelho et al., 2016); [10] (Dœlsch et al., 2006); [11] (Ahn and Chon, 2010); [12] (Brtnický et al., 2020); [13] (Cabral Pinto et al., 2015); [14] (Amaral et al., 2006); [15] (Mora et al., 2012); [16] (Mendoza-Grim´ on et al., 2014); [17] (Sutherland and Tack, 2000); [18] (Kazakou et al., 2010). F. Afonso et al. Chemosphere 334 (2023) 138973 8 4.1. Metal loadings in water and sediment In Praia das Conchas and Diogo Nunes despite being systems with high metal content, there was no contamination for most of the assessed metals and are, simultaneously, represented by areas with small amounts of sediment organic matter. Due to the strong affinity between organic matter and metals, there is less ability for the sediment to trap and store the metals available in the system. Although these systems had a high metal content in water and sediment samples, the indices did not show contamination for most of the metals assessed. Contrarily, Malanza was the system with the highest proportions of silt and organic matter in the sediment, which increases the number of covalent bonds with metals, favouring their accumulation in the sediment. Hence, these sites could have an impact on the element retention in the sediment, even though this phenomenon alone cannot explain the contamination and enrichment in the system. Table 3 Indices estimated for each metal in samples from all study areas (EFEnrichment Factor; Igeo – Geoacumulation index; CF – Contamination Factor). F. Afonso et al. Chemosphere 334 (2023) 138973 9 4.2. Natural and anthropogenic sources of metal enrichment and contamination To better understand the natural tendency for metal concentration, the collected data was compared with the results from sediment samples from other mangroves in the tropical region and with volcanic islands. Comparison between sediments from the study area with other mangrove systems has shown Santomean mangroves to have higher metal contents. However, land systems from other volcanic islands presented higher copper, lead and nickel concentrations, although the background values (OSPAR Commision, 2009) for copper and nickel are lower than the values detected in S˜ ao Tom´ e. This suggests that the higher content in copper and nickel might have a lithogenic origin. Nevertheless, the background values considered are restrictively representative of the North-East Atlantic Region which means that conclusions based on these values must be taken with caution due to different geological formations. A common observation in all systems was the high enrichment and contamination with arsenic and chromium. In non-volcanic islands, the contamination with arsenic has been proven to be derived from agriculture runoff and the use of disinfection and pest elimination detergents in aquaculture practices, while the chromium contamination was considered to be mainly a result of agriculture runoff (Li et al., 2022). In volcanic islands, the arsenic contamination in mangrove systems resulted from natural processes, such as the leaching of volcanic rocks, but also from anthropogenic origin, linked to harbours and wastewater discharge. The same type of contamination with arsenic was found in mangroves from the mainland possibly caused by anthropogenic activities, for instance, petrochemical wastes and desalination plants (Al-Kahtany et al., 2018). However, chromium contamination is most commonly the result of freshwater input and erosion of volcanic rocks (Cary, 1982). In Praia das Conchas and Diogo Nunes, the sources of enrichment in iron and zinc are most probably local and short-term, due to the anthropogenic activities that are common in the system (i.e., laundry washing and trash deposition). Diogo Nunes, however, showed a contamination with chromium and manganese, which can result from a long-term release of these metals, possibly from agriculture run-off, the main land use in the area. Angolares presented the lowest concentration for most of the metals and fewer signs of sediment contamination out of all studied mangroves. Contrary to what was expected, considering the proximity of the landfill from the south arm, it was the northern arm that presented higher metal contamination. This may be explained by either the proximity to the city of S˜ ao Jo˜ ao dos Angolares, or by a higher metal contribution of that particular hydrographic basin. In Malanza, besides the possible sources of pollution from the river tributaries, this system may also have on other possible external source of contamination, the I´ ogoi´ ogo bay, due to intense fishing and other anthropogenic activities and the regular boat traffic. For instance, near the harbour, the concentration of copper, lead, and zinc was clearly higher, and these are metals usually present in boat emissions and antifouling chemicals used in boat paints (Badr et al., 2009; Naser, 2013; Usman et al., 2013; Zhang et al., 2012). This may also justify the different metals present in the water column of the Malanza mangrove from the upstream and downstream areas, as a result of seawater intrusion from the nearby bay, as the harbour is close to the mangrove. Sediments can be used to assess the environmental condition of a specific ecosystem, based on enrichment and contamination indexes that evaluate the sample and metal load. However, the use of several indices can result in irregular outputs and inconsistencies between them. For instance, the Igeo scale can range from negative to positive values, while all the other indices have different scales that only range in positive values, being ambiguous how they complement each other and which classes are equivalent (Birch, 2023). Thus, even though classification schemes for sediments can be an advantage to any study related to the evaluation of sediment quality and metal pollution, the outputs must be considered with caution. 4.3. Health implications In developing countries, urbanization had a rapid advance in the last years, however, the lack of management plans associated with the excessive use of fertilizers and disposal of domestic waste in inappropriate areas have caused detrimental effects on the quality of water, sediment, and aquatic fauna (Ali et al., 2016; Isa et al., 2014; Ledin et al., 1989). Additionally, the countries’ legislation about ecosystem contamination is weak and the lack of monetary support compromises the actual law enforcement (Kumie and Kloos, 2006). As a consequence of some of those limitations, it was created the One Health approach (www.onehealthcommision.org) focused on protecting people’s health, animals, and the environment. This approach addresses the root causes and searches for sustainable solutions applicable to food and water safety, as well as pollution management. In this scope, and considering the specific case of S˜ ao Tom´ e mangroves, some of the possible responses to prevent more health damages by metal contamination would be the implementation of policies for ecosystem monitoring (i.e., monitoring metal concentration in shellfish and fish from local markets), as well as the implementation of antior low-pollution policies. The development and investment in education programs are also important tools to raise awareness on the best practices for system protection, pollution sources and impacts on the ecosystems and human health (Beyene et al., 2009). Furthermore, engaging with the communities can help reduce health risk increasing practices, as well as to detect early symptoms of metal contamination. 5. Conclusions To the best of our knowledge, this was the first assessment of the metal concentration in the mangroves from S˜ ao Tom´ e. It showed that these systems present higher metal concentrations than most mangroves and volcanic islands, probably due to the combination of a volcanic origin, which tends to increase the natural metal concentrations than sites with other lithogenic origins, and sediment composition that favours the retention of those metals. However, occasional high concentrations and high contamination values suggest that, at least in some cases, the high concentration of some metals have an anthropogenic origin. Arsenic and chromium showed severe enrichment and contamination at different sites, which raises health concerns. These results reinforce the need to gain a better understanding of actual sources of metal contamination in S˜ ao Tom´ e, which can then enable the preparation and application of mitigation measures. Future studies and monitoring of metal concentrations, enrichment and contamination in S˜ ao Tom´ e should go beyond mangrove areas, extending to other aquatic and terrestrial ecosystems, and including the study of bioaccumulation. Overall, this study shows how metal pollution might be relevant even in a remote and non-industrialized island, highlighting the need for a better understanding of how metals are distributed in the environment, and particularly in developing countries, where people are often more directly reliant on natural resources and more likely to unknowingly be exposed to high metal concentrations and its deleterious health consequences. Credit author statement Filipa Afonso: Investigation, Writing – original draft, Visualization; Carla Palma: Validation, Formal analysis, Resources, Writing – review & editing, Visualization; Ana C. Brito: Writing – review & editing; Paula Chainho: Investigation, Writing – review & editing; Ricardo de Lima: Investigation, Writing – review & editing, Visualization; Joshua A. Heumüller: Investigation, Writing – review & editing; Filipe Ribeiro: Investigation, Writing – review & editing; Pedro M. F´ elix: F. Afonso et al.