https://orcid.org/0000-0002-7025-8598
Figures
Abstract
Flooding compromises the microbiological quality of domestic and recreational water sources, increasing exposure to waterborne and opportunistic pathogens in affected communities. Following widespread flooding triggered by the September 2023 controlled spillage of the Akosombo and Kpong dams in Ghana, which led to massive flooding, we assessed the microbial quality of water bodies in communities downstream of the Volta River and associated public and ecosystem health implications. In July 2024, water was collected from multiple source types across four downstream townships (Asutsuare, Aveyime, Battor, and Mepe), including river sections, ponds, puddles, wells, boreholes, and canal-associated sites. Samples were filtered, microbial DNA was extracted, and bacterial community composition was profiled using 16S rRNA gene sequencing, with downstream bioinformatics analyses used to characterize microbial diversity, assess spatial variation across sites, and stratify detected taxa according to ecological origin and public health risk. Sequence data obtained from 17 out of 22 samples revealed communities dominated by Proteobacteria with frequent co-occurrence of Bacteroidetes and Firmicutes. Across all sites, 56 microbial taxa and 118 microorganisms were detected. Community composition varied by location and water source type, with bank and mid-river sites generally exhibiting higher richness than puddles and a canal drain. Thirteen samples contained site-specific taxa, indicating marked spatial heterogeneity within and between communities. Microorganisms of public health relevance were frequently detected, including enteric-associated bacteria such as Escherichia coli, Salmonella enterica, Shigella species, and Cronobacter sakazakii, as well as opportunistic and healthcare-associated bacteria including Acinetobacter baumannii, Klebsiella pneumoniae, Pseudomonas aeruginosa, and Stenotrophomonas maltophilia. Higher-risk organisms were most widely distributed in Mepe and Battor, including in groundwater-associated sources. These findings provide molecular evidence of persistent, spatially heterogeneous microbial hazards months after flooding, underscoring continued exposure risks in downstream communities and the need for targeted water safety interventions and risk-informed public health responses.
Citation: Kpeli GS, Agyirey-Kwakye P, Livingstone CN, Cudjoe LT, Dotse EE, Nyarko E, et al. (2026) Downstream community risks post dam-spillage flooding along the Volta River in Ghana: potential pathogens and public health implications. PLoS One 21(7): e0346766. https://doi.org/10.1371/journal.pone.0346766
Editor: Marcos Pileggi, Universidade Estadual de Ponta Grossa, BRAZIL
Received: March 23, 2026; Accepted: June 16, 2026; Published: July 2, 2026
Copyright: © 2026 Kpeli et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All raw sequencing data generated in this study have been deposited in the NCBI Sequence Read Archive (SRA) under BioProject accession number PRJNA1469279. GPS coordinates for all sampling sites are provided in Supporting Information S1 Table. Data underlying all figures and analyses, including microbial diversity and abundance charts, Krona plots, alpha diversity indices and plots, and microbial risk classification data, are provided in Supporting Information S1 Fig, S2 Fig, and S2 Table. Laboratory and analytical protocols are publicly available on protocols.io under the following DOIs: • Environmental Water Sample Collection Following Flooding: https://dx.doi.org/10.17504/protocols.io.4r3l2dd64g1y/v2 • Vacuum Filtration and DNA Extraction from Environmental Water Samples: https://dx.doi.org/10.17504/protocols.io.kqdg3mmkpl25/v2 • 16S rRNA Amplicon Sequencing Using Ion GeneStudio S5: https://dx.doi.org/10.17504/protocols.io.5jyl84468g2w/v1 • Bioinformatics and Ecological Risk Classification of Environmental Water Microbiomes: https://dx.doi.org/10.17504/protocols.io.rm7vz44o2lx1/v1.
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Events that occur post-flooding are a major concern for public health, stemming from the fact that there is the drastic deterioration of drinking and recreational water quality due to increased microbial contamination ([1,2]). Floodwaters act as vectors, transporting a diverse range of microbial pathogens and fecal contaminants from various sources, including agriculture, industrial effluents, human and animal waste, into receiving rivers and streams. This poses dire, long-lasting health challenges to affected communities ([3,4]). An increase in microbial contamination implies an increase in the risk of acquiring waterborne diseases such as dysentery, typhoid fever, leptospirosis, diarrhea etc. which may be life-threatening especially for the vulnerable (children, elderly individuals, immunocompromised individuals, pregnant women etc.) members of affected communities [5,6]. An additional post-flood concern is that affected individuals are at risk of developing skin, ear, eye and respiratory infections due to the presence of diverse pathogens [7].
These infections, depending on their severity and repeated occurrences, may lead to chronic, life-long organ damage, examples being bladder cancer and liver fibrosis caused by Schistosomiasis infection. In addition to the risk of drowning from rising water levels, infections caused by pathogens found in waterbodies post-flood may lead to increased rates of mortality [8]. Studies have consistently demonstrated a rise in fecal indicator bacteria in flood-affected rivers, emphasizing the increased likelihood of microbial contamination and subsequent health risks [9]. The presence of indicator organisms such as Escherichia coli, fecal coliforms, enterococci, Pseudomonas aeruginosa, Clostridium perfringens, and specific viruses and protozoans, signals potential fecal contamination and the increased risk of pathogenic microorganisms known for causing diseases such as diarrhea, urinary tract infections (UTIs), cholera among many others [10,11]. These contaminants pose significant health threats to communities that depend heavily on affected water sources for survival (domestic purposes) frequently leading to outbreaks of waterborne diseases [12,13].
Beyond the immediate public health concerns, the broader impacts of flooding extend to socio-economic consequences, including reduced agricultural yields and harsh effects on the health-related quality of life for affected populations, manifesting as increased rates of depression, anxiety, and chronic pain that can persist for years [14,15].
The Kpong Dam, a significant hydroelectric facility on Ghana's Volta River, was constructed between 1977 and 1982 and since its construction, has been extremely beneficial to Ghana through the generation of electricity [16–18]. In addition to power generation, the dam serves as a water source for irrigation and municipal water supply [19]. However, its operation poses serious public health challenges for downstream communities, with Bilharzia (Schistosomiasis) endemic in some of these communities since its construction [20–22]. Facing critical water levels due to unusually heavy rainfall in late 2023, the Volta River Authority (VRA) as part of regulatory duties initiated a controlled spillage with increased discharge volumes on September 15th 2023, to avert potential structural failure. This controlled spillage coinciding with heavy rainfalls at the time, resulted in widespread and devastating flooding across downstream communities along the Volta River highly impacting 7 districts of the Volta, Eastern and Greater Accra Regions [23,24]. Towns including Mepe, Sogakope, Battor, Adidome, Mafi, Tefle, Wume and Sokpoe were severely hit, with Mepe experiencing the worst inundation. The spillage triggered a humanitarian crisis, displacing tens of thousands of individuals, leading to significant losses of homes, livelihoods, and critical public health infrastructure [25,26].
Understanding the microbiological quality of water post-flood particularly in downstream communities is necessary for assessing the extent of contamination, identifying potential health threats, and implementing effective risk mitigation strategies and public health interventions. This paper aimed at assessing the potential emergence of pathogenic and potentially pathogenic bacteria post flooding, with implications for human and ecosystem health. We sought to comprehensively investigate the complex and varying effects of the Kpong Dam spillage on public health in the downstream communities of Asutsuare, Aveyime, Battor and Mepe communities using 16S rRNA gene sequencing of water samples from these towns.
Materials and methods
The study utilized a cross-sectional approach to assess the presence and prevalence of pathogenic and potentially pathogenic bacteria in the downstream communities with implications for human health post-flooding. Water samples were collected from four townships; Asutsuare, Aveyime, Battor, and Mepe, located downstream of the Kpong dam in the North Tongu District of the Volta Region. These communities except for Asutsuare were affected by the controlled spillage from the dam. Sample collection was done in July 2024, when floodwaters had receded, and it was safe to go into the communities. The interval between the September 2023 dam spillage and the July 2024 sampling reflects two sequential phases: an initial sampling round was conducted once floodwaters had receded and it was safe to access the communities, but DNA quality from those initial samples was insufficient for sequencing. A second sampling round was therefore undertaken in July 2024 to obtain samples of adequate quality. Importantly, this extended post-flood window also allowed assessment of the persistence of microbial contamination well after the acute flooding phase. All study procedures, including field site access and sample collection, were conducted under ethical clearance obtained from the Research Ethics Committee of the University of Health and Allied Sciences (UHAS), Ho. Approval was granted for each study community under the following protocol identification numbers: Asutsuare (UHAS-REC A.10 [162] 23–24), Aveyime (UHAS-REC A.10 [245] 23–24), Battor (UHAS-REC A.10 [248] 23–24), and Mepe (UHAS-REC A.10 [175] 23–24). No additional permits were required for field site access, as sampling was conducted from publicly accessible water bodies along the Volta River downstream of the Kpong Dam.
Sampling design and sample collection
Sampling sites were selected in the field to represent various water sources within the community, including sections of the Volta River, taps, wells, boreholes, ponds, and other local water bodies. The exact locations of each sampling site were geo-referenced using a handheld GPS device, and coordinates were recorded to ensure spatial traceability. GPS coordinates for all 22 sampling sites are provided as supporting information S1 Table. Water samples were collected using the direct immersion grab sampling method. Sterile bottles were immersed below the surface, and the caps were removed only after the bottles were submerged to minimize surface contamination and ensure representative water collection. At each site, 1.5 L of water was collected in sterile containers in accordance with Environmental Protection Agency (EPA) guidelines [27] for water sample collection. Each sample was given a unique identifier and transported on ice, in tightly sealed sterile styrofoam boxes. For quality control, samples were taken in duplicate from each site. A total of 22 samples were obtained from the Volta River, as well as other sources in the townships including wells, domestic taps, ponds, and puddles. Except for the Battor township where 7 samples were collected, 5 samples each were collected in the Asutsuare, Aveyime, and Mepe townships. In the Asutsuare township: 1 sample was obtained from a canal, 1 from the canal drain and 3 from different sections across the river. From Aveyime: 2 samples were obtained from the river, 1 from a pond, and 2 from two separate puddles, on the riverbank. From Battor: 4 samples were obtained from the river, and one each from a pond, borehole, and a domestic groundwater well. From Mepe: 1 sample was obtained from a pond; 3 obtained directly from the river (banks and middle), while one was collected from a household groundwater well. The detailed sampling protocol is available at https://dx.doi.org/10.17504/protocols.io.4r3l2dd64g1y/v2 [28].
Sample processing
Samples underwent vacuum filtration using a Whatman 0.45µm filter paper (Cytvia, UK). The filter papers were subsequently immersed in 1X phosphate-buffered saline solution to resuspend the cells. DNA was extracted using the QIAamp® DNA Mini Kit (Qiagen N.V., Germany), following the manufacturer’s instructions. The extracted DNA was stored at −20oC till use. The concentrations and purities of the extracts were assessed using a NanoDropTM Spectrophotometer (Thermo Fisher Scientific, U.S.) at the Noguchi Memorial Institute for Medical Research, Accra, Ghana; and QubitTM dsDNA HS quantification (Thermo Fisher Scientific, Waltham, Massachusetts, U.S.) at AttoDiagnostics Limited, Norwich, UK. The detailed laboratory protocol is available https://dx.doi.org/10.17504/protocols.io.kqdg3mmkpl25/v2 [29].
16S rRNA gene sequencing and post-sequencing analyses
16S rRNA gene sequencing was performed for 17 samples which were made up of 2 samples from Asutsuare, 5 from Aveyime, 6 from Battor, and 4 from Mepe. Five other samples were not processed due to low DNA quality. DNA sequencing was done at AttoDiagnostics Limited, Norwich, UK. 16S rRNA amplicons covering 8 out of the 9 hypervariable regions (HV2 to HV9) were amplified using the Ion AmpliSeq Microbiome Health Research Kit, followed by targeted sequencing using the Ion 540 Chip (Cat# 46496, ThermoFisher Scientific, USA) on the Ion GeneStudio™ S5 platform (Thermo Fisher Scientific, U.S.) following the manufacturers protocol. Demultiplexing and quality control processes were performed using Ion ReporterTM Software version 5.18 (Thermo Fisher Scientific, U.S.). Taxonomic assignments were made using the Greengenes database (version 13.5), while alpha diversity metrics were calculated and visualized using the RStudio IDE version 2025.05.0 + 496 (“Mariposa Orchid”) (Posit Software, PBC). The minimum read counts for the different taxonomic levels were set at 500 for the genus and family, and 100 for the species level, while normalization thresholds were established at 0.5 for genus and family levels, and 0.25 for the species. Results obtained after the sequencing and post-sequencing analyses were organized using Microsoft Excel LTSC Standard Version 16.99.1 (25072013). The detailed sequencing protocol is available at: https://dx.doi.org/10.17504/protocols.io.5jyl84468g2w/v1 [30].
Ecological and risk-based classification of microorganisms
Following 16S rRNA sequencing, microbial isolates were categorized using an ecological and risk-based classification framework into healthcare-associated, environmental/aquatic, enteric and opportunistic organisms. This enabled the grouping of microorganisms by both their likely origin and potential impact on human health. Healthcare-associated microbes included organisms commonly found in healthcare settings and typically associated with nosocomial infections; environmental and aquatic microbes referred to those naturally occurring in soil, water bodies and surrounding ecosystems without direct association with human disease. Enteric organisms were defined as those originating from faecal contamination and typically associated with the gastrointestinal tract of humans and animals while opportunistic pathogens included microbes which are generally harmless in healthy individuals but have the potential to cause infection in the immunocompromised. This classification facilitated the understanding of the microbial diversity present in the samples collected while highlighting potential health risks associated with different microbes.
Data analysis
All data was organized in Microsoft Excel Version 16.99.1 (25072013) and analyzed using RStudio (IDE version 2025.05.0 + 496 “Mariposa Orchid”; Posit Software, PBC) and the R packages ggplot2, dplyr, tidyr, readr, viridis, and pheatmap. The detailed bioinformatics and risk classification protocol is available at https://dx.doi.org/10.17504/protocols.io.rm7vz44o2lx1/v1 [31].
Results
Taxonomic composition and diversity of microbial isolates
From the 17 samples successfully sequenced, a total of 56 distinct microbial taxa and 118 distinct microorganisms were recorded (Table 1). Samples from all 4 towns were consistently dominated by Proteobacteria, whiles Bacteroidetes and Firmicutes also frequently co-occurred across sites. Acinetobacter were detected at lower relative abundances in several samples
At the town level, samples from Asutsuare comprised 34 microbial taxa with the bank sample (ALBQ) showing the highest richness (28 taxa) and a strong dominance of Proteobacteria (89%). Multiple enteric and opportunistic organisms such as Escherichia coli, Salmonella enterica, Shigella flexneri, Shigella sonnei, Acinetobacter baumannii, Klebsiella pneumoniae, Pseudomonas aeruginosa, and Stenotrophomonas maltophilia were detected, whereas the canal drain sample (CDQ) exhibited lower diversity (6 taxa) and was largely composed of environmental and opportunistic organisms including Bacteroides sp., Comamonas testosteroni, Sphingomonas sp., and Desulfovibrio sp.
In Aveyime, a total of 31 taxa were identified across five sampling sites. Proteobacteria remained the dominant phylum in all the samples (64–89%), with variable representation of Bacteroidetes and Firmicutes. The bank and mid-river samples showed higher taxonomic richness compared with puddle samples, with several opportunistic Gram-negative species repeatedly detected. Detected organisms included environmental and opportunistic Gram-negative bacteria such as Pseudomonas putida, Flavobacterium sp., Acidovorax sp., and Comamonas aquatica, as well as potentially pathogenic taxa including Escherichia coli, Aeromonas hydrophila, Aeromonas veronii, and Burkholderia cepacia.
Samples from Battor showed the highest overall diversity, with 38 microbial taxa identified. Phylum-level distributions were more heterogeneous, with notable contributions from Firmicutes and Actinobacteria alongside Proteobacteria. Detected taxa included opportunistic Gram-negative organisms such as Acinetobacter baumannii, Pseudomonas aeruginosa, and Stenotrophomonas maltophilia, as well as Gram-positive and anaerobic taxa including Bacillus sp., Clostridium sp., Staphylococcus sp., and Streptococcus sp. Borehole and bank samples showed broad taxonomic representation, including anaerobic, environmental, and opportunistic organisms compared with the pond and mid-river samples.
In Mepe, 36 taxa were identified, with substantial variation across sites. The pond sample demonstrated complete dominance of Proteobacteria (100%) and included multiple enteric species such as Escherichia coli, Salmonella enterica, Shigella boydii, Shigella flexneri, and Klebsiella sp, while the bank and mid-river samples showed higher overall richness and broader phylum representation. Several samples displayed extensive representation of Acinetobacter species (A. baumannii, A. lwoffii, A. calcoaceticus) alongside other opportunistic taxa such as Pseudomonas sp., Legionella sp., and Stenotrophomonas maltophilia (Table 1).
All 17 successfully sequenced samples displayed distinct microbial profiles. Of these, 13 samples contained microbial taxa that were not detected at any other sampling site indicating the presence of site-specific microorganisms. These unique taxa were distributed across all townships and multiple sample types, including bank, pond, mid-river, canal drain, borehole, and well samples. Several sampling sites were characterized by single unique taxa, such as Desulfovibrio sp. in the canal drain sample (CDQ), Chitinimonas taiwanensis in AV. A, Alcaligenes faecalis in BP, and Bdellovibrio bacteriovorus in the Mepe bank sample. Other sites recorded multiple unique microorganisms, including ALBQ, BB, BW, and M. Pond, each of which harbored between two and five taxa not identified elsewhere. The unique microorganisms identified spanned a range of ecological groups, including enteric bacteria (e.g., Shigella sonnei, Shigella boydii, Enterobacter asburiae), environmental and aquatic taxa (e.g., Flavobacterium psychrophilum, Flavobacterium columnare, Bdellovibrio bacteriovorus), and opportunistic or healthcare-associated organisms (e.g., Acinetobacter nosocomialis, Acinetobacter junii, Chryseobacterium anthropi). Together, these findings highlight pronounced site-level heterogeneity in microbial composition across the sampled environments (Table 1).
Across all the towns, bank and mid-river samples generally exhibited higher microbial richness compared with puddle and canal drain samples. Unique taxa were detected at multiple sites, indicating localized microbial signatures within and between communities. There were nine unique microbes identified in the Aveyime Township, none of which were detected in the other townships. Similarly, 5, 16, and 13 unique microbes were identified in Asutsuare, Battor and Mepe respectively. The UpSet plot (Fig 1) further illustrates the distribution and overlap of microbial taxa across the four townships, highlighting both shared microbial communities and township-specific taxa.
The left horizontal bar chart (A) represents the total number of unique microbial taxa identified within each individual township. The main vertical bar chart (B), together with the intersection matrix (C), shows the number of microbial taxa present exclusively in specific township combinations, as indicated by dark dots, and absent from the remaining townships.
Microbial community structure and beta diversity
Principal Coordinates Analysis (PCoA) was used to assess beta diversity and examine differences in microbial community composition across sampling sites and townships (Fig 2). The first two principal coordinate axes explained 21.57% (PCoA Axis 1) and 17.57% (PCoA Axis 2) of the total variation, capturing the major gradients in community dissimilarity among samples. Distinct clustering and dispersion patterns were observed across towns with samples from Battor clustering within the lower right quadrant of the ordination space, with relatively uniform spacing, indicating a moderately consistent microbial community structure across sites. In contrast, Aveyime samples exhibited greater dispersion, reflecting higher within-site variability in community composition. Mepe samples were heterogeneously distributed, with bank and mid-river samples clustering together yet clearly separated from other sites within the town, suggesting spatial heterogeneity in microbial communities within the township. Samples from Asutsuare displayed the greatest degree of dispersion in ordination space, indicating pronounced variation in microbial community composition among sites. Notably, the Asutsuare bank sample (ALBQ) clustered closely with samples AV.1 (Aveyime) and M. Pond (Mepe), suggesting partial overlap in microbial community structure across townships. This overlap may reflect shared environmental characteristics or hydrological connectivity influencing microbial assemblages at these sites.
Each point represents a unique sampling site with points coloured by townships (Mepe, Battor, Aveyime, Asutsuare). The distance between points reflects the degree of dissimilarity in the microbial composition. The plot uses two main axes, PCoA Axis 1 (explaining 21.57% of the variation) and PCoA Axis 2 (explaining 17.57% of the variation), to capture the largest dimensions of microbial community differences.
Risk-based stratification of microorganisms with ecological subclassification
The identified microorganisms were classified both by risk levels and ecological origin. Microorganisms were stratified broadly by risk levels into high, medium, and low categories. Within each risk level, organisms were further categorized according to their ecological classification, such that healthcare-associated pathogens, environmental/aquatic organisms, enteric pathogens, and opportunistic organisms were represented as subsets of the respective risk categories (Table 2).
The spatial frequency and site-level distribution of pathogenic and potentially pathogenic taxa are illustrated in Fig 3 providing spatial context to the observed risk profiles. The figure demonstrates that clinically relevant microorganisms were not uniformly distributed across sites but instead showed clear site-specific and hydrology-linked patterns. Bank, mid-river, and pond environments exhibited higher frequencies of detection of pathogenic species compared with canal drain and puddle sites, Multiple pathogenic taxa were detected at more than one site within the same town, while other taxa were restricted to single sampling locations.
The figure illustrates the detection patterns of clinically relevant microorganisms classified as high- or medium-risk across sampling sites. Each point represents the presence of a pathogenic or potentially pathogenic taxon at a given site. The distribution highlights site-specific and repeated detection patterns of organisms of public health importance across towns and hydrological settings.
Variation in detection frequency was observed across towns, with differences in the number of sites at which pathogenic taxa were detected. Sites exhibiting higher clustering of pathogenic taxa corresponded to towns with greater representation of high- and medium-risk organisms in the risk-based classification (Table 2).
Town-level distribution of risk-classified microorganisms
Asutsuare.
Microbes were found across all risk levels and spanned most of the ecological categories. High-risk microbes found included healthcare-associated pathogens such as Acinetobacter baumanii, Klebsiella oxytoca, Klebsiella pneumoniae, and Serratia marcescens, enteric pathogens such as Escherichia coli, Salmonella enterica, Shigella flexneri, and Cronobacter sakazakii; and Legionella sp. as the only environmental/aquatic organism. No high-risk opportunistic microbes were detected. Medium-risk organisms included healthcare-associated pathogens such as Citrobacter freundii, and multiple Enterobacter species; and enteric pathogens including Aeromonas hydrophila, and Bacillus cereus. Opportunistic pathogens included Pseudomonas putida, Pantoea agglomerans, Burkholderia sp., Shingomonas sp. and Comamonas testosteroni and no medium-risk environmental/aquatic microbes were found. Low-risk organisms included Propionibacterium acnes as the single healthcare-associated pathogen; Flavobacterium sp., Iodobacter sp., Acidovorax sp., and Desulfovibrio sp. as environmental/aquatic microbes; and Rickettsia endosymbiont as the single opportunistic pathogen. No low-risk enteric microbes were detected.
Aveyime.
Samples from the town harboured a high burden of medium-risk opportunistic taxa. The high-risk microbes detected included healthcare-associated pathogens: Acinetobacter baumannii, Pseudomonas aeruginosa, Stenotrophomonas maltophilia, and Burkholderia cepacia. Escherichia coli and S. maltophilia were the sole enteric and opportunistic pathogens identified respectively. No high-risk environmental/aquatic microbes were detected. Medium-risk organisms detected include the healthcare-associated pathogens: Acinetobacter sp., Alcaligenes sp., and Stenotrophomonas sp. while multiple Flavobacterium species were identified as environmental/aquatic microbes. The Enteric pathogens detected were Aeromonas hydrophila, Aeromonas veronii, Bacillus sp., and Chitinimonas taiwanensis and opportunistic pathogens included Alcaligenes sp., Bacillus sp., and multiple Comamonas and Pseudomonas species. In the low-risk category, Brevundimonas sp. and Rickettsia endosymbiont were the sole low-risk healthcare-associated and opportunistic pathogens identified respectively. Acidovorax sp., and Enhydrobacter aerosaccus were detected as low-risk environmental/aquatic microbes and no low-risk enteric pathogens were identified.
Battor.
Battor showed the widest distribution of microorganisms across all risk levels and ecological categories. High-risk healthcare-associated pathogens such as Acinetobacter baumanii, Burkholderia cepacia, Klebsiella oxytoca, and Serratia marcescens were identified whiles Legionella sp. was the only high-risk environmental/aquatic organism detected. The Enteric pathogens detected included Escherichia coli, Shigella sp. and Salmonella enterica, while opportunistic pathogens identified were Burkholderia cepacia and Stenotrophomonas sp. Medium-risk organisms included healthcare-associated pathogens, such as multiple Acinetobacter species, Citrobacter freundii, and Propionibacterium acnes, as well as environmental/aquatic taxa including Pseudomonas sp., Rhodococcus sp., and Thermomonas haemolytica. This risk category also comprised potentially pathogenic enteric organisms, such as multiple Aeromonas species and Bacillus cereus, together with opportunistic pathogens including Burkholderia sp., Chryseobacterium anthropi, Chryseobacterium hominis, Alcaligenes faecalis, and multiple Pseudomonas species. Low-risk healthcare-associated pathogens detected included Acinetobacter beijerinckii, Acinetobacter lwoffii, and Acinetobacter radioresistens. Low-risk environmental/aquatic organisms comprised multiple Flavobacterium species and Bdellovibrio bacteriovorus, while Sphingobacterium sp. was the only low-risk opportunistic pathogen identified. No enteric pathogens were detected within this risk category.
Mepe.
Microorganisms detected in Mepe spanned all three risk levels and several ecological categories. High-risk organisms included healthcare-associated pathogens such as Acinetobacter baumannii, Pseudomonas aeruginosa, and Burkholderia cepacia. Legionella sp. was detected as the sole high-risk environmental/aquatic organism. High-risk enteric pathogens comprised Escherichia coli, Salmonella enterica, and multiple Shigella species, while Stenotrophomonas maltophilia and Burkholderia cepacia were the only high-risk opportunistic pathogens identified. Medium-risk organisms included healthcare-associated pathogens such as multiple Acinetobacter and Enterobacter species, as well as Klebsiella sp. Environmental/aquatic organisms detected at this risk level included Flavisolibacter sp. and multiple Flavobacterium species. The enteric pathogens identified were Aeromonas hydrophila and Aeromonas veronii, while opportunistic pathogens comprised multiple Pseudomonas, Rhodococcus, and Enterobacter species. Low-risk organisms included healthcare-associated taxa such as Acinetobacter sp. and Pseudomonas putida. Environmental/aquatic organisms detected at this risk level included Pectobacterium carotovorum, Bdellovibrio bacteriovorus, and Thermomonas haemolytica. No low-risk enteric or opportunistic pathogens were detected.
Discussion
This study provides molecular evidence of microbial contamination and potential public health risk in downstream communities along the Volta River following the 2023 controlled spillage of the Akosombo and Kpong dams. Using 16S rRNA gene sequencing, we identified diverse microbial communities across multiple water sources in Asutsuare, Aveyime, Battor, and Mepe, including enteric pathogens, opportunistic organisms, and healthcare-associated bacteria of established public health relevance. Recent evidence from a longitudinal study has shown that flood-impacted aquatic environments can remain microbiologically compromised for extended periods after floodwaters recede, particularly in low and middle-income settings where sanitation and water infrastructure are disrupted [4]. The findings from this study are consistent with these observations and highlight the persistence of microbial hazards in downstream environments months after flooding events.
Across all the towns, microbial communities were dominated by Proteobacteria, with frequent co-occurrence of Bacteroidetes and Firmicutes. Similar taxonomic patterns have been reported in flood-affected rivers and disturbed freshwater systems, where increased nutrient loading and hydrological disturbance favour Proteobacteria-rich assemblages [32,33]. In this study, bank and mid-river samples generally exhibited higher microbial richness than puddle and canal drain samples. Comparable richness patterns have been documented in recent flood and riverine microbiome studies, where hydrologically connected environments support more diverse microbial communities than isolated or stagnant waters [34,35]. For downstream communities that rely on riverbanks and mid-channel waters for domestic activities, this increased diversity includes a broader spectrum of organisms with potential public health implications [36].
Beta diversity analysis further demonstrated pronounced spatial heterogeneity in microbial community composition across sampling sites and towns. Partial clustering by town was observed alongside substantial within-town dispersion, particularly in Asutsuare and Mepe. Similar patterns of spatial heterogeneity have been described in aquatic microbial ecology studies, where local environmental conditions, water source type, and contamination inputs shape microbial communities over short spatial scales [37]. The overlap observed between selected sites across different towns likely reflects shared hydrological connectivity within the Volta River system, a feature that has been noted in large river basins affected by dam operations and controlled spillages [38].
A key public health signal in this study is the repeated detection of enteric-associated organisms, including Escherichia coli, Salmonella enterica, Shigella sp., and Cronobacter sakazakii, across multiple towns and water source types. Floodwaters are known to mobilize faecal contaminants from sanitation facilities, agricultural areas, and animal waste into surface and groundwater sources, thereby increasing the risk of waterborne disease transmission [4,39]. Recent studies in flood-affected settings have consistently linked elevated levels of enteric bacteria with increased incidence of diarrhoeal and gastrointestinal diseases, particularly among children and other vulnerable populations [40]. The strong enteric signal observed in Mepe’s pond sample is therefore of particular concern given the documented severity of flooding and displacement in this community.
In addition to enteric pathogens, this study identified multiple opportunistic and healthcare-associated organisms of clinical relevance, including Acinetobacter baumannii, Klebsiella pneumoniae, Pseudomonas aeruginosa, and Stenotrophomonas maltophilia. Recent environmental and public health research increasingly recognizes natural and engineered water systems as reservoirs for such organisms, especially in post-flood contexts where infrastructure damage and altered water flow facilitate environmental dissemination [4,41]. The detection of these taxa in groundwater-associated sources such as wells and boreholes, particularly in Battor, aligns with reports that shallow or poorly protected groundwater sources are vulnerable to contamination following flooding events [42].
The integration of ecological classification with risk-based stratification provides a structured framework for interpreting microbial findings in a public health context. Similar risk-oriented approaches have been recommended in recent environmental health literature which advocates adapting Quantitative Microbial Risk Assessment (QMRA) frameworks for rural and remote community water systems by incorporating source-specific ecological heterogeneity to enable tiered risk prioritization beyond presence-absence reporting [43]. The pathogen-focused spatial analysis in this study demonstrates that pathogenic and potentially pathogenic taxa are not uniformly distributed across sampling sites, with repeated detection at certain locations and restriction to single sites at others. Such spatial heterogeneity has been widely reported in flood-affected water systems and underscores the importance of site-specific surveillance and intervention rather than uniform community-wide risk assumptions [44–47].
Our study also presents with a few limitations. Detection of bacterial DNA using 16S rRNA sequencing does not confirm organism viability or infectivity, and species-level resolution may be limited for some taxa. The cross-sectional design, conducted after floodwaters had receded, does not capture temporal dynamics during the acute flooding phase. In addition, antimicrobial resistance profiles were not assessed, despite the detection of organisms commonly associated with multidrug resistance. Nonetheless, the combined taxonomic, beta diversity, and risk-based analyses provide a comprehensive snapshot of microbial hazards in downstream communities following dam-related flooding. Future studies should prioritize longitudinal sampling across seasons, quantitative microbial assays, culture-based confirmation, and resistance gene profiling, to better characterize ongoing public health risks in downstream flood-affected settings.
Overall, our study highlights that downstream flood-affected communities along the Volta River system are exposed to spatially heterogeneous microbial risks that vary by water source type and location. The persistence of enteric, opportunistic, and healthcare-associated bacteria months after flooding underscores the need for targeted water safety interventions, strengthened environmental surveillance, and context-specific public health responses in communities impacted by dam spillages and extreme hydrological events.
Supporting information
S1 Fig. Microbial diversity, abundance and community composition charts.
Graphical summaries of microbial relative abundance and community composition across sampling sites, including Krona plots illustrating the taxonomic distribution of microbial communities detected in each sample.
https://doi.org/10.1371/journal.pone.0346766.s001
(DOCX)
S2 Fig. Alpha diversity indices and plots.
Tabulated alpha diversity indices and associated graphical representations of within-sample microbial diversity across all sampling sites and townships.
https://doi.org/10.1371/journal.pone.0346766.s002
(DOCX)
S1 Table. GPS coordinates of all 22 sampling sites.
Geographic coordinates recorded for each sampling site across the four study townships (Asutsuare, Aveyime, Battor, and Mepe) using a handheld GPS device.
https://doi.org/10.1371/journal.pone.0346766.s003
(DOCX)
S2 Table. Microbial risk classification data.
Tabulated data detailing the township, sampling site, identified microorganism, risk level, ecological category, group classification, and associated disease for all microbial taxa detected across sampling sites.
https://doi.org/10.1371/journal.pone.0346766.s004
(XLSX)
Acknowledgments
We acknowledge the School of Basic and Biomedical Sciences, University of Health and Allied Sciences for allowing us to use their labs for this work. We would also like to acknowledge Prof Anthony Ablordey and Dr Prince Asare of the Noguchi Memorial Institute for Medical Research for their assistance.
References
- 1. Talbot CJ, Bennett EM, Cassell K, Hanes DM, Minor EC, Paerl H, et al. The impact of flooding on aquatic ecosystem services. Biogeochemistry. 2018;141(3):439–61. pmid:30930510
- 2. Kamalanandhini M, Kalaivizhi R, Golda Percy VP, Srividhya S, Dheepak S, Thiyaneshwaran KK. Effect of Flood Event on Water Quality. Rasayan Journal of Chemistry. 2019;12(02):849–54.
- 3. Collender PA, Cooke OC, Bryant LD, Kjeldsen TR, Remais JV. Estimating the microbiological risks associated with inland flood events: Bridging theory and models of pathogen transport. Crit Rev Environ Sci Technol. 2016;46(23–24):1787–833. pmid:28757789
- 4. Dzodzomenyo M, Asamoah M, Li C, Kichana E, Wright J. Impact of flooding on microbiological contamination of domestic water sources: a longitudinal study in northern Ghana. Applied Water Science. 2022;12(10):235.
- 5. Haldar K, Kujawa-Roeleveld K, Hofstra N, Datta DK, Rijnaarts H. Microbial contamination in surface water and potential health risks for peri-urban farmers of the Bengal delta. Int J Hyg Environ Health. 2022;244:114002. pmid:35759864
- 6. Cabral JPS. Water Microbiology. Bacterial Pathogens and Water. Int J Environ Res Public Health. 2010;7(10):3657–703.
- 7. Huang L-Y, Wang Y-C, Wu C-C, Chen Y-C, Huang Y-L. Risk of Flood-Related Diseases of Eyes, Skin and Gastrointestinal Tract in Taiwan: A Retrospective Cohort Study. PLoS One. 2016;11(5):e0155166. pmid:27171415
- 8. Lynch VD, Sullivan JA, Flores AB, Xie X, Aggarwal S, Nethery RC, et al. Large floods drive changes in cause-specific mortality in the United States. Nat Med. 2025;31(2):663–71. pmid:39753964
- 9. Mukherjee M, Hossain MS, Boswell J, Zhang Y, Allen MS, LaMontagne MG, et al. Large, but short-term, increase in fecal indicator bacteria following extreme flooding from Hurricane Harvey in Houston, TX. Frontiers in Water. 2024;6:1466377.
- 10. Myers DT, Zgleszewski L, Bier R, Price JR, Morgan S, Jackson JK. High fecal indicator bacteria in temperate headwater streams at baseflow: implications for management and public health. Water Res X. 2025;29:100347.
- 11. Rochelle-Newall E, Nguyen TMH, Le TPQ, Sengtaheuanghoung O, Ribolzi O. A short review of fecal indicator bacteria in tropical aquatic ecosystems: knowledge gaps and future directions. Front Microbiol. 2015;6:308. pmid:25941519
- 12. Acosta-España JD, Romero-Alvarez D, Luna C, Rodriguez-Morales AJ. Infectious disease outbreaks in the wake of natural flood disasters: global patterns and local implications. Infez Med. 2024;32(4):451–62. pmid:39660153
- 13. Yeboah SIIK, Antwi-Agyei P, Kabo-Bah AT, Ackerson NOB. Water, environment, and health nexus: understanding the risk factors for waterborne diseases in communities along the Tano River Basin, Ghana. J Water Health. 2024;22(8):1556–77. pmid:39212287
- 14. Abass K, Gyasi RM, Katey D, Frempong F, Garsonu EK. Flood exposure and psychological distress among Ghanaian adults in flood-prone settings. Sci Total Environ. 2022;835:155481. pmid:35490804
- 15. Othman F, Sharip SM, Kasim S, Kamdari MN, Baharuddin FN. Flooding: The mental, social and health effects of flood disasters. 2022.
- 16.
World Bank. Ghana: Volta River Authority (VRA); appraisal of the Kpong Hydroelectric Project. World Bank. 1977. https://documents1.worldbank.org/curated/en/571301468282896201/pdf/multi-page.pdf
- 17.
Natural Heritage Institute, Water Resources Commission G, Volta River Authority, Partners. Project Summary: Reoptimization and Reoperation Study of Akosombo and Kpong Dams. 2014.
- 18.
Volta River Authority. Volta River Authority | Kpong Hydro Plant. https://www.vra.com/our_mandate/kpong_hydro_plant.php 2025. Accessed 2025 July 22.
- 19. Dadzie H. Water supply to communities affected by spillage of Akosombo and Kpong dams will not be contaminated by debris- GWCL. GBC Ghana Online. https://www.gbcghanaonline.com/general/water-supply/2023/ 2023. Accessed 2025 July 22.
- 20. Yirenya-Tawiah DR, Rashid AA, Futagbi G, Aboagye I, Dade M. Prevalence of snail vectors of schistosomiasis in the Kpong head pond, Ghana. West African Journal of Applied Ecology. 2011;18:40–5.
- 21. Zakhary K. Factors Affecting the Prevalence of Schistosomiasis in the Volta Region of Ghana. MJM. 2020;3(2):93–101.
- 22. Opare J, Hervie T, Mensah E, Brown-Davies C, Asiedu O, Alomatu B, et al. Schistosomiasis in Ghana from baseline to now: the impact of fifteen years of interventions. Front Public Health. 2025;13:1554069. pmid:40547465
- 23. Smallwood T, Lefebvre V, Eyre R, Harrison J, Delaporte S, Attandoh R. Ghana, Akosombo floods: Long- and medium-distance displaced subscribers from flood-affected areas of the Lower Volta River: Situation as of 22 October 2023. 2023. https://www.flowminder.org/resources/publications-reports/ghana-lower-volta-river-akosombo-floods-october-2023#
- 24.
Volta River Authority. Volta River Authority | News - Controlled Spillage of Water from the Akosombo Dam and Kpong Dam. https://www.vra.com/media/2023/Controlled-Spillage-from-the-Akosombo-Dam-and-Kpong-Dam.php 2023. Accessed 2024 January 30.
- 25.
Akosombo Dam spillage: Everyone downstream will die if the dam overflows – Former VRA CEO rubbishes any blame. Modern Ghana. https://www.modernghana.com/news/1402101/akosombo-dam-spillage-everyone-downstream-will.html 2025. Accessed 2025 July 22.
- 26. Arku SK. All you need to know about Ghana’s Akosombo Dam spillage. Peoples Dispatch. https://peoplesdispatch.org/2023/11/01/all-you-need-to-know-about-ghanas-akosombo-dam-spillage/ 2023. Accessed 2025 July 22.
- 27.
Ohio Environmental Protection Agency. Appendix IV-Inland Lakes Manual. Surface Water Field Sampling Manual. 2025.
- 28. Kpeli GS, Agyirey-Kwakye P, Livingstone CN, Cudjoe LT, Dotse EE, Nyarko E. Environmental water sample collection following flooding. protocols.io. https://dx.doi.org/10.17504/protocols.io.4r3l2dd64g1y/v2 2026.
- 29. Kpeli GS, Agyirey-Kwakye P, Livingstone CN, Cudjoe LT, Dotse EE, Nyarko E. Vacuum Filtration and DNA Extraction from Environmental Water Samples. https://dx.doi.org/10.17504/protocols.io.kqdg3mmkpl25/v2 2026.
- 30. Kpeli GS, Agyirey-Kwakye P, Nutifafa Livingstone C, Teye Cudjoe L, Edem Dotse E, Nyarko E. 16S rRNA amplicon sequencing using Ion GeneStudio S5 v1. protocols.io. 2026.
- 31. Kpeli GS, Agyirey-Kwakye P, Nutifafa Livingstone C, Teye Cudjoe L, Edem Dotse E, Nyarko E. Bioinformatics and ecological risk classification of environmental water microbiomes v1. 2026.
- 32. Ting ASY, Zoqratt MZHM, Tan HS, Hermawan AA, Talei A, Khu ST. Bacterial and eukaryotic microbial communities in urban water systems profiled via Illumina MiSeq platform. 3 Biotech. 2021;11(2):40. pmid:33479595
- 33. Zheng F, Zhang T, Yin S, Qin G, Chen J, Zhang J, et al. Comparison and interpretation of freshwater bacterial structure and interactions with organic to nutrient imbalances in restored wetlands. Front Microbiol. 2022;13:946537. pmid:36212857
- 34. Rahman T, Haider SMA, Khan FTZ, Islam MS, Mahmud ZH, Karim MM, et al. Flood-Driven Microbial Resurgence: Functional Shifts and Beneficial Taxa in Post-Flood Agricultural and Residential Soils in Bangladesh. Microbiologyopen. 2025;14(5):e70069. pmid:41085228
- 35. Zeglin LH. Stream microbial diversity in response to environmental changes: review and synthesis of existing research. Front Microbiol. 2015;6:454. pmid:26042102
- 36. Duze ST, Mkhize L, Marimani M, Patel M. A river in crisis: water quality, microbial burden, and public health implications of a South African urban river. Appl Environ Microbiol. 2025;91(11):e0156625. pmid:41060022
- 37. Xiang H, Yang C-G, Xiong H-Q, Bao H-R, Qu J-Z, Luan Z-X, et al. Spatial heterogeneity of microbial community structure and its environmental drivers in surface sediments of Erhai Lake. PLoS One. 2025;20(9):e0322915. pmid:40906763
- 38. Chen H, Zhang L, Zheng Z, Gao Y, Zhao Y. Hydrological connectivity shape the nitrogen pollution sources and microbial community structure in a river-lake connected system. Front Microbiol. 2025;16:1563578. pmid:40291805
- 39. Tan B, De Vera P, Abrazaldo J, Ng C. Flood-associated disease outbreaks and transmission in Southeast Asia. Front Microbiol. 2025;16:1694246. pmid:41199949
- 40. Yazdi MS, Ardalan MA, Hosseini M, Yousefi Zoshk M, Hami Z, Heidari R, et al. Infectious Diarrhea Risks as a Public Health Emergency in Floods; a Systematic Review and Meta-Analysis. Arch Acad Emerg Med. 2024;12(1):e46. pmid:38962364
- 41. Hayward C, Ross KE, Brown MH, Bentham R, Whiley H. The Presence of Opportunistic Premise Plumbing Pathogens in Residential Buildings: A Literature Review. Water. 2022;14(7):1129.
- 42. Krishnakumar A, Jose J, Kaliraj S, Aditya SK, Krishnan KA. Assessment of the impact of flood on groundwater hydrochemistry and its suitability for drinking and irrigation in the River Periyar Lower Basin, India. Environ Sci Pollut Res Int. 2022;29(19):28267–306. pmid:34988810
- 43. Rodrigues PM. Quantitative microbial risk assessment in rural and remote communities: a critical review. IOSR Journal of Environmental Science, Toxicology and Food Technology. 2025;19:67–76.
- 44.
World Health Organization. Quantitative Microbial Risk Assessment: Application for Water Safety Management. https://iris.who.int/handle/10665/246195 2016. Accessed 2026 January 12.
- 45. Watkin S, Ciric L, Kiernan M, Cloutman-Green E. A proposed classification system for opportunistic pathogens for improved healthcare infection prevention and control risk assessments. J Hosp Infect. 2023;135:206–7. pmid:36948226
- 46. Gerba CP. Environmentally Transmitted Pathogens. In: Pepper IL, Gerba CP, editors. Environmental Microbiology: Third Edition. Elsevier Inc. 2014:509–50.
- 47. Denissen J, Reyneke B, Waso-Reyneke M, Havenga B, Barnard T, Khan S, et al. Prevalence of ESKAPE pathogens in the environment: Antibiotic resistance status, community-acquired infection and risk to human health. Int J Hyg Environ Health. 2022;244:114006. pmid:35841823