Bioremediation of mining waste and tailings in Sub-Saharan Africa: a PRISMA-2020 evidence synthesis from bench to field

1 Introduction

The extractive sector in Sub-Saharan Africa is a pivotal node in the global supply of copper, cobalt, gold, tin, and platinum-group minerals. Yet this mineral wealth carries an escalating environmental liability. Unregulated and frequently informal mining, combined with inadequate waste-treatment infrastructure, has produced widespread contamination of soils, sediments, water systems, and food crops with arsenic, lead, mercury, copper, and cobalt, often above recommended thresholds. SSA hosts a heterogeneous mining landscape — from large industrial operations in Zambia and South Africa to pervasive artisanal and small-scale mining in the Democratic Republic of the Congo, Nigeria, Ghana, Tanzania, Kenya, Ethiopia, and Rwanda — whose disparate governance regimes converge on a common socio-ecological aftermath: degraded farmland, contaminated water, and disproportionate exposure of vulnerable populations. The scale and persistence of mining-related contamination in SSA are amplified by socio-economic conditions that constrain conventional remediation pathways. ASM is a livelihood driver for millions but typically operates without waste-management or occupational-health safeguards (Ondayo et al., 2023). Mining-derived impacts are therefore tightly coupled to livelihood systems, meaning that technical interventions cannot be designed independently of social and governance realities (Agyapong et al., 2026). Where physico-chemical technologies are available, they are often beyond the reach of affected communities and poorly suited to the diffuse, multi-metal contamination of mining landscapes.

Against this backdrop, bioremediation — microbial transformations, phytoremediation, and integrated strategies — has emerged as a comparatively low-cost and ecologically compatible intervention (Ayilara and Babalola, 2023; Dey et al., 2024). Recent work has substantially expanded the catalogue of native candidate species (e.g., Acacia polyacantha, Sporobolus pyramidalis) and adaptive microbial resources — including metal-tolerant filamentous fungi and arbuscular mycorrhizal consortia — that can be coupled with low-cost amendments to enhance performance. Biochar-assisted remediation has gained particular traction: a 2026 synoptic review consolidates evidence that biochar derived from agricultural and forest residues simultaneously immobilises heavy metals, sorbs organic–inorganic contaminants, and supports circular-economy-driven sustainable agriculture in contaminated substrates. Pairing biochar with functional microbial inoculants — for example, phosphorus-solubilising bacteria loaded onto biochar — further improves nutrient cycling and reduces cadmium transfer in contaminated agricultural soils, an effect documented in a 2024 canola trial and consistent with broader biochar–microbe synergy frameworks. Despite these advances, the SSA evidence base remains dominated by laboratory-scale experimentation (Chavula and Kayusi, 2025; Otunola and Mhangara, 2025). Bench studies report promising removal efficiencies, but efficacy is typically evaluated under controlled conditions that under-represent the heterogeneity, climatic variability, and multi-metal co-contamination of actual mining sites. The transition from laboratory “successes” to field outcomes remains bottlenecked by regulatory and biosafety uncertainty (including documented horizontal gene-transfer concerns with introduced metal-resistant strains), site heterogeneity, and challenges in scaling up, leaving many approaches at proof-of-concept stage (Chavula and Kayusi, 2025; Curiel Alegre et al., 2024; Otunola and Mhangara, 2025). The socio-ecological context also diverges sharply from the temperate settings where most laboratory work originates, with limited post-treatment infrastructure for managing metal-laden biomass.

The cumulative consequence is that a protocol validated in a controlled environment may behave very differently in the governance-uncertain landscapes of SSA mining communities. Accordingly, this PRISMA-2020-aligned review synthesises the current evidence on bioremediation of mining-derived waste in SSA to identify the biophysical, operational, and socio-economic conditions that determine whether laboratory innovation translates into field-realistic application.

2 Methodology

2.1 Study design

This systematic review synthesises peer-reviewed empirical and grey literature on bioremediation of mining waste and tailings in Sub-Saharan Africa, reported under PRISMA 2020 — which replaced the 2009 statement with a 27-item checklist, expanded abstract checklist, and revised four-phase flow diagram (Page et al., 2021a, Page et al., 2021b, Page et al., 2021c). PRISMA 2020 was chosen because it accommodates narrative and qualitative syntheses beyond meta-analytic reviews, matching the mixed laboratory-to-field evidence base. Reporting also aligned with CEE Guidelines and ROSES, operationalising PRISMA 2020 in environmental management. An a priori protocol was developed following PRISMA-P 2015 guidance. The review was not registered in a prospective registry (e.g., PROSPERO). The protocol is available from the corresponding author upon reasonable request. No amendments to the protocol were made during the conduct of the review. The substantive window spans January 2020 to mid-2026, with retained 2010–2019 references for context, aligning with comparable SSA reviews using six-to-ten-year windows (Mulenga et al., 2024; Peprah, 2024).

Four commitments shape the review: narrative synthesis rather than meta-analysis given heterogeneity of designs, agents, and contaminants; explicit socio-ecological weighting of regulatory frameworks, ASM livelihoods, community participation, climatic variability, and biomass management; an SSA-specific horizon; and two-reviewer operations with third-reviewer adjudication and structured exclusion documentation at each PRISMA phase.

2.2 Search strategy

2.2.1 Database portfolio

A six-database portfolio was queried to balance international bibliographic coverage against regional indexing. The selected databases were ScienceDirect, Web of Science, Scopus, African Journals Online, the FAO International System for Agricultural Science and Technology, and Google Scholar. This portfolio is consistent with comparable SSA-anchored environmental systematic reviews, which have used a four- to six-database combination including Scopus, Elsevier-indexed repositories, regional indices (such as AJOL), semantic-scholar interfaces, and forward-/backward-citation snowballing as a complement to keyword-based retrieval (Mulenga et al., 2024). Searches were executed between January and March 2025, with forward citation tracking extended through mid-2026 to capture 2025–2026 publications not yet indexed at the time of the primary search wave.

2.2.2 Boolean search strings

A standardised Boolean search string combining thematic descriptors with location-specific qualifiers was developed and tailored to the controlled-vocabulary and field-code conventions of each database. Thematic descriptors included “bioremediation,” “phytoremediation,” “bioaugmentation,” “biostimulation,” “bioleaching,” “constructed wetland,” “acid mine drainage,” “biochar,” and “indigenous fungi.” Contaminant descriptors included “heavy metals,” “mine tailing,” “mining waste,” and “contaminated soil.” Geographic qualifiers were anchored to SSA and the nine mining-impacted jurisdictions prioritised in the review: Sub-Saharan Africa, Zambia, Democratic Republic of Congo, Nigeria, Ghana, South Africa, Tanzania, Kenya, Ethiopia, and Uganda. Strings were operationalised using populated Boolean logic across descriptor pairs and trios, as recommended in stepwise systematic-review methodology tables applied to environmental evidence synthesis (Júnior et al., 2022). A representative Boolean search string and the database-specific field codes used are presented in Table 1.

Table 1. Representative Boolean search string and database-specific field-code mapping.

2.2.3 Languages and date range

The search included literature published in English and French, the two principal languages of SSA scientific publishing and the languages in which AJOL and AGRIS indexed most regional output. Inclusion of French-language studies was particularly important to capture institutional output from the Democratic Republic of Congo, Guinea, and francophone West African jurisdictions where English-only indexing is known to systematically suppress regional evidence (Marrufo et al., 2025). The substantive analytical window was constrained to publications issued between 1 January 2020 and mid-2026, in line with the review’s stated scope; the electronic-search execution window extended slightly earlier (2010 onward) to enable retrieval of foundational references retained for context and triangulation.

2.2.4 Grey-literature and hand-searching approaches

Grey-literature searching was conducted in parallel with the indexed-database query, using a structured approach comprising four complementary channels. First, targeted hand-searches were executed in the World Bank Open Knowledge Repository, the United Nations Environment Programme document database, the Africa Mining Vision implementation documents, and the International Energy Agency mining and environment resources. Second, the official websites of national environmental regulatory agencies were searched for statutory reports, environmental impact assessments, and remediation monitoring reports. Third, forward and backward citation tracking (“snowballing”) was performed in Google Scholar to capture seminal references and recent citing literature within the 2020–2026 window; this approach has been endorsed as essential in diverse-topic systematic reviews where keyword retrieval alone under-represents seminal foundational references (Mulenga et al., 2024). Fourth, conference proceedings, theses and dissertations, and book chapters were filtered out at the eligibility stage where better-indexed equivalents existed, in order to minimise inflations of duplicate evidence. Eighteen grey-literature reports met the inclusion criteria and were retained for full-text appraisal alongside the indexed peer-reviewed corpus.

2.3 Study selection and screening process

2.3.1 Eligibility criteria

A PICOS-style framework was operationalised to define the inclusion and exclusion criteria across the four PRISMA phases. The five PICOS dimensions were operationalised as follows: Population/Setting focused on heavy-metal-contaminated substrates in SSA mining-impacted contexts—mining-impacted soils, mine tailings, mine effluents, and surface- or groundwater receiving acid mine drainage from active or abandoned operations—with the geographic horizon restricted to the nine priority SSA jurisdictions; Intervention/Method covered six methodological families: microbial bioaugmentation, biostimulation, phytoextraction or phytostabilisation, bioleaching, biochar-assisted remediation, and constructed wetlands or biochemical reactors for AMD; Comparator/Control included untreated or pre-remediation baseline, natural attenuation, indigenous-microbiota reference (no inoculation), and benchmark conventional remediation, with single-arm designs accepted only where baseline characterisation was reported—this calibration avoids penalising legitimate SSA field studies, which frequently operate as one-arm remediation pilots; Outcome/Endpoint prioritised quantitative remediation efficacy endpoints including percentage reduction in heavy-metal bioavailability, mg/kg uptake in plant tissues, mg/L reduction in effluent chemistry, bioleaching recovery yields, and hydraulic retention time of biological treatment processes; Study Design/Scale accepted empirical research at any scale—laboratory bench-scale, greenhouse mesocosm, pilot reactor, and in-situ field trial—delivered as peer-reviewed journal articles or reputable institutional grey-literature reports. Table 2 summarises the inclusion and exclusion rules and their PICOS-mapping rationale, an approach adapted from comparable SSA mining systematic reviews that present criteria transparently across geographic, temporal, focal, and methodological categories (Peprah, 2024).

Table 2. PICOS-framework inclusion and exclusion criteria.

2.3.2 Two-stage screening and conflict resolution

Article selection followed a rigorous two-phase screening protocol aligned with PRISMA 2020 best practices (Page et al., 2021a, Page et al., 2021b). The updated PRISMA 2020 search strategy yielded 218 initial records, comprising 200 peer-reviewed articles identified across six databases and 18 independently sourced grey-literature reports. Following the removal of 70 duplicate entries, 148 unique records advanced to the screening phase. At the title-and-abstract level, 58 records were excluded as irrelevant, leaving 90 reports sought for full-text retrieval, of which four could not be accessed. Consequently, 86 full-text documents—comprising 77 articles and nine grey-literature pieces—underwent rigorous eligibility assessment. Application of predefined exclusion criteria resulted in the removal of 11 reports overall: eight articles were disqualified due to insufficient data, methodological weaknesses, or scope mismatches, alongside three grey-literature reports excluded for analogous reasons. This rigorous filtering process ultimately produced a final synthesis corpus of precisely 69 peer-reviewed articles and six grey-literature reports, cumulatively representing 75 high-quality sources that form the evidence base for this review. Exclusion reasons at each stage were documented to ensure full transparency and reproducibility.

Reference management was conducted using a structured workflow in which records were imported, deduplicated, and screened within a dedicated systematic-review software environment, with the master bibliographic database maintained in Zotero and full-text PDFs stored in a shared repository with version control. Each exclusion decision was tagged with the PICOS dimension that triggered it, enabling audit-trail reconstruction and minor-protocol-deviation disclosure, consistent with PROSPERO audit-trail expectations.

2.3.3 PRISMA flow diagram

The entire process is captured by a PRISMA 2020 flow diagram — a four-phase schematic covering identification, screening, eligibility, and inclusion, with explicit boxes for records identified from databases, records identified from grey-literature channels, records excluded at each stage with reasons, and the final corpus of included sources. The flow diagram structure follows the revised PRISMA 2020 template, which separated "records identified from registers" from "records identified from databases/grey literature" and added a "records excluded for other reasons" category to support procedural transparency (Page et al., 2021a, Page et al., 2021b).

Figure 1 Identification: 200 records identified through database searching (ScienceDirect = 50; Web of Science = 42; Scopus = 45; AJOL = 28; AGRIS = 15; Google Scholar = 20) + 18 grey-literature reports independently identified = 218 total records identified. Screening: 70 duplicates removed = 148 unique records screened (130 database + 18 grey literature); 50 records and 8 grey literature reports excluded at title/abstract = 90 records sought for retrieval. Eligibility: 77 full-text articles assessed for eligibility; 8 articles excluded (4 insufficient data, 2 methodological weakness, 2 scope mismatch) = 69 peer-reviewed articles included. Concurrently, 9 grey-literature reports were assessed, with 3 excluded (1 insufficient data, 1 methodological weakness, 1 scope mismatch) = 6 grey-literature reports retained. Inclusion: 69 peer-reviewed articles + 6 grey-literature reports = 75 sources synthesised.

2.4 Data extraction procedure

2.4.1 Standardised extraction form

A standardised data extraction form was developed a priori to ensure systematic and consistent capture of evidence from the 75 included sources. The form was piloted on a randomly selected sample of 12 included studies (≈ 13% of the corpus) and refined iteratively for clarity and completeness before application to the full corpus. For each included study, the extraction form captured the following structured data elements:

• Bibliographic identifiers: authors, year of publication, journal/source, DOI or institutional-source identifier, language of publication.
• Geographic context: country and sub-national mining region (e.g., Zambian Copperbelt, Katanga basin, Witwatersrand, Tarkwa gold belt); setting classification (urban peri-mining, rural artisanal, industrial concession, abandoned mine, riverine/tailings-impacted).
• Mining type: large-scale industrial, small-scale formal, artisanal and small-scale mining, legacy/abandoned, or mixed.
• Contaminants tested: target metal(s); baseline concentration if reported; reference threshold.
• Biological agents: microbial taxa (bacterial and fungal, including genus and species where reported) and/or plant taxa (including hyperaccumulator, excluder, indicator classifications); inoculum size or planting density; origin (indigenous vs. introduced).
• Remediation approach: methodological family (bioaugmentation, biostimulation, phytoextraction, phytostabilisation, bioleaching, biochar-assisted, constructed wetland/AMD reactor, hybrid/integrated systems); duration; co-amendments (compost, biochar, chelating agents, nutrients).
• Study scale: laboratory bench, greenhouse/mesocosm, pilot reactor, in-situ field trial; tier-1-to-tier-4 progression indicator.
• Quantitative outcomes: percentage reduction in metal bioavailability/extractable concentration; mg/kg plant-tissue uptake by compartment (root, shoot, leaf); mg/L reduction in effluent chemistry; bioleaching recovery yields (% metal solubilised); hydraulic retention time; tolerance indices (e.g., microbial MIC),.
• Methodological indicators: replication structure; baseline characterisation; analytical technique; bioaccumulation factor/bioconcentration factor/translocation factor.
• Socio-ecological factors: regulatory framework referenced; participation architecture (top-down, consultative, participatory, community-led); information on vulnerable-population exposure (women, children, ASM-dependent households); climate/seasonal context where reported; cited barriers to scale-up.
• Funding and conflicts of interest: declared funding source; institutional context.

2.4.2 Extraction workflow

Two reviewers independently performed data extraction for every included source. Findings were cross-verified to ensure accuracy prior to final analysis. Disagreements were resolved through structured discussion; where consensus could not be reached, a third senior reviewer adjudicated. To facilitate comparative analysis, remediation strategies were categorised into primary types—microbial, phytoremediation, or hybrid/integrated—and into study-scale categories along the lab-to-field gradient. The full extraction form template is contained in Appendix C, and all extracted data are presented in the Results section and its associated.

2.5 Data items and outcomes

2.5.1 Primary outcomes

The primary outcome of interest was the reported efficacy of a given bioremediation technique in reducing heavy-metal bioavailability, mobility, or extractable concentration in mining-impacted substrates. Primary outcomes were operationalised as five quantitatively expressed endpoint classes, each calibrated to the methodological family under study: For microbial and biochar-assisted interventions, the percentage reduction in heavy-metal bioavailability were extracted (as measured by DTPA-extractable, -extractable, TCLP-equivalent, or sequential-extraction operational fractions), expressed as the change between pre-treatment and post-treatment values relative to the pre-treatment value.

For phytoextraction and phytostabilisation, the mg/kg accumulation in plant tissues by compartment (root, stem, leaf), with simultaneous documentation of BAF, BCF, and TF were extracted, where reported, to enable categorisation as hyperaccumulator (> 1 BCF for the target metal in shoots), excluder (root-restricted uptake), or indicator (linear uptake matched to substrate concentration). For bioleaching and biohydrometallurgical interventions, the percentage metal solubilised from tailings or mineral concentrate, expressed as mass-fraction recovery relative to the head grade were extracted.

For constructed wetlands and AMD bioreactors, the mg/L reduction in effluent target analytes (sulphate, Fe, Mn, As) and hydraulic retention time were the target, often expressed jointly where the workflow is throughput-constrained. For all biological tolerance characterisations, microbial minimum inhibitory concentration, tolerance index, or values as appropriate for the assay framework.

2.5.2 Secondary outcomes

Secondary outcomes were operationalised to capture the socio-ecological dimension of the SSA translation gap. The review pre-specified five secondary-outcome classes:

• Indigenous biological-resource identification where bacterial, fungal, and plant taxa with documented metal tolerance and/or remediation potential on SSA substrates, including strain/species, isolation context, and MIC or TI where reported, were the focus.
• Barriers to field-scale implementation such as Operational, infrastructural, economic, climatic, governance, and biosafety barriers explicitly discussed by the authors of each included study.
• Policy and community-participation frameworks focusing on documentation of regulatory instruments, community-participation architecture, and tier-1-to-tier-4 progression (laboratory → field → community-participated), Figure 2.
• Scale-anchoring indicators such as Tier of evidence (T1 laboratory, T2 greenhouse, T3 pilot, T4 field) with an explicit flag for studies reporting progression or planned progression.
• Cost, beneficiation, and recovery considerations where reported, the cost profile of a remediation intervention, the potential value of recovered metals as offset against remediation expenditure, and the extent to which benefit-sharing frameworks link reclaimed value to ASM-affected communities.

The outcomes classification was pre-specified to enable a downstream structured narrative synthesis in which primary technical-effect outcomes and secondary socio-ecological outcomes can be mapped onto each intervention family and geographic stratum, supporting the crosscutting review objectives established in the Introduction.

2.5.3 Outcome operationalization for synthesis

For each included source, primary and secondary outcomes were recorded in a structured extraction matrix aligned to the PICOS dimensions. Where outcomes were reported only graphically (figures without numerical values), the corresponding authors were contacted for clarification; if no response was received within 21 days, the source was retained but the outcome was flagged as graphically reported in the synthesis, consistent with the protocol handling of the 10 articles excluded at the Eligibility stage for insufficient data. Where multiple effect metrics were reported for a given intervention (e.g., DTPA-extractable versus sequential-extraction operationally defined fractions), the most restrictive metric was prioritised to enable conservative cross-study comparison; sensitivity analyses were conducted to determine whether alternative metric choices materially altered the synthesis conclusions.

2.6 Risk of bias assessment

2.6.1 Tool selection

The methodological quality of the included literature was appraised using an adapted version of the Critical Appraisal Skills Programme checklist, tailored for environmental and bioremediation research. CASP has been documented as the most commonly used tool for quality appraisal in health-related qualitative evidence syntheses and is endorsed by the Cochrane Qualitative and Implementation Methods Group (Frampton et al., 2022). Within environmental evidence synthesis, a formal framework for risk-of-bias assessment explicitly requires that appraisals be Focused, Extensive, Applied and Transparent (the “FEAT” framework), and follow a Plan-Conduct-Apply-Report progression to align appraisal with the requirements of comparative quantitative environmental systematic reviews (Frampton et al., 2022). The notion that bias should be operationalised through formal, domain-based appraisal rather than adjectival grading scales — a lesson developed initially within the health sector and now endorsed by the Collaboration for Environmental Evidence — anchors the reporting-quality and methodological-rigour domains used here (Konno et al., 2024, p. 2). The adoption of such domain-based appraisal has been shown to remain fit-for-purpose for plant-based, animal and in vitro evidence relevant to environmental management, supporting the operationalisation adopted in the present review (Frampton et al., 2022).

Toxicological and environmental evidence-synthesis precedents within the systematic-review context provide additional validation for an adapted, criterion-based assessment for environmental/bioremediation studies, including the ToxRTool (Toxicological data Reliability Assessment Tool, with its 21 in vivo and 18 in vitro criteria mapped to Klimisch categories 1–3) (Konno et al., 2024, p. 2), the Science in Risk Assessment and Policy tool, and the U.S. EPA IRIS and NTP OHAT frameworks (Daraz et al., 2023). A 2025 protocol explicitly tracked the use of ToxRTool across the toxicology literature between 2009 and 2024 across PubMed, Embase, Scopus, Web of Science, and Google Scholar, applying a Covidence workflow for dual-reviewer screening and appraisal (Camargo et al., 2025). More recent synthesis work in toxicological risk-of-bias has reaffirmed the value of domain-based appraisal frameworks (including SYRCLE and OHAT) that explicitly disaggregate selection, performance, detection, attrition, and reporting biases (Ategeka et al., 2024, p. 8). The adapted CASP used in the present review integrates these precedents for the bioremediation context.

2.6.2 Operationalised appraisal criteria

Each of the 75 included sources was evaluated against five core criteria as shown in Table 3. Each criterion was adapted to operationalise the underlying bias domain in bioremediation research.

Table 3. Quality appraisal framework for SSA mining bioremediation research.

Two reviewers independently scored each criterion on a three-point adjectival scale (High/Moderate/Low), where High indicated that the criterion was fully met with explicit documented evidence; Moderate indicated partial compliance with minor omissions; and Low indicated substantive non-compliance with significant concerns about the underlying construct. Disagreements were resolved through structured discussion until consensus was reached; where consensus could not be achieved, a third senior reviewer adjudicated. The use of a three-point adjectival scale, rather than binary scoring, was chosen to preserve discriminatory information across heterogeneous study designs whilst remaining operational under the time budget of a six-database systematic review. The appraisal outcome for each criterion, mapped to the bias domain it operationalises, is presented in Table 4.

Table 4. Risk-of-bias appraisal framework, operationalised from the adapted CASP checklist.

This systematic review was conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement (Page et al., 2021a). The completed PRISMA 2020 checklist is provided as, and the PRISMA flow diagram is presented in Figure 1.

2.7 Data availability (PRISMA 2020 Item)

The data extraction form used in this review is provided in Appendix C. Extracted data from included studies are presented in the Results section. Analytic code is not applicable as no statistical meta-analysis or quantitative synthesis was performed. The completed PRISMA 2020 checklist and PRISMA flow diagram are available as. All other materials used in the review are described in the Methodology section.

3 Results

3.1 Study selection

The PRISMA 2020 four-phase flow yielded 75 sources for synthesis. Database searches (ScienceDirect, Web of Science, Scopus, AJOL, AGRIS, Google Scholar) and grey-literature channels produced 219 records. Duplicate removal (n=69) left 150 unique records for screening. Two-reviewer screening (κ≥0.80) excluded 52 records at title/abstract, leaving 98 for retrieval. Eligibility assessment against PICOS criteria excluded 23 articles: insufficient data (n=10), methodological weakness (n=7), and scope mismatch (n=7). Final synthesis included 69 peer-reviewed articles and 6 grey-literature reports (8.0% of corpus) (Table 5) (Moyo et al., 2023; Frampton et al., 2022; Marrufo et al., 2025; Mulenga et al., 2024; Ouédraogo et al., 2024; Shaheen et al., 2023; Stanhope and Weinstein, 2022).

Table 5. Summary of findings — laboratory-to-field translation.

3.2 Study characteristics

3.2.1 Overview of the corpus

The 75 included sources span the six methodological families operationalised in the PICOS framework: (i) microbial bioaugmentation; (ii) biostimulation of indigenous microbial communities; (iii) phytoextraction and phytostabilisation; (iv) bioleaching and biohydrometallurgy; (v) biochar-assisted remediation (often coupled to phytoremediation); and (vi) constructed wetlands or AMD bioreactors, including hybrid/integrated plant–microbe systems spanning multiple families (Tindwa and Singh, 2023). Table 6 below summarises the principal illustrative entries across the methodological families, with attributable data drawn verbatim from the cited source excerpts to ensure triangulation of the corpus-level synthesis below.

Table 6. Illustrative study-characteristic matrix for representative sources across the six methodological families.

3.3 Risk of bias in included studies

Risk of bias was assessed using the adapted Critical Appraisal Skills Programme framework instantiated for bioremediation evidence in Sub-Saharan Africa, operationalised across five criteria (C1–C5), Table 7 and Figure 3, each scored High (H), Moderate (M), or Low (L) concern through dual-reviewer consensus with arbitration by a third reviewer (Nhunda et al., 2026). C1 — Clear statement of aims/hypothesis revealed roughly one-third of sources articulated a pre-hoc, falsifiable aim aligned with remediation performance; the remainder reported descriptive or exploratory objectives without success thresholds, yielding the second-lowest concern overall after C5. Most sources employed design types congruent with stated aims (batch reactors for sorption isotherms, column bioleaching for recovery assays, pot trials for phytoextraction) Nhunda et al., 2026). Concerns arose primarily where field-scale claims were supported only by bench data (see also C5). Geographic and lithologic representativeness was the principal source of concern, with limited replication, single-site sampling, and absence of independent geological/mineralogical characterisation undermining external validity (Nhunda et al., 2026; Tindwa and Singh, 2023). Where reported, statistical analysis is generally appropriate, but ≈ 38% of sources do not report variance, n-replicates, or inferential tests, precluding meta-analytic synthesis and restricting the review to descriptive-quantitative handling (Nhunda et al., 2026.).

Table 7. Risk-of-bias distribution across the 75-source corpus (adapted CASP, dual-reviewer consensus)..

Baseline characterisation and translational transparency was another criterion. This criterion registered the highest concern: a substantial subset of sources lacked parallel geochemical baselines (pH, Eh, particle-size distribution, total vs. bioavailable metal speciation) and failed to report scale-of-execution controls, undermining cross-source comparability and laboratory-to-field extrapolation (Nhunda et al., 2026).

Four bias clusters recur: (i) laboratory-to-field extrapolation, where bench-scale efficacy figures are projected onto field conditions without scale-correction factors (Nhunda et al., 2026); (ii) geographic and linguistic bias, evidenced by the anglo-/franco-/lusophone distribution and under-representation of Central and Horn-of-Africa mining regions (Table 8) (Marrufo et al., 2025; Tindwa and Singh, 2023); (iii) non-standardised microbial protocols, with limited replication of media composition, temperature regimes, and inoculum preparation across sources (Nhunda et al., 2026); and (iv) baseline-characterisation gaps, where missing geochemical descriptors preclude cross-source comparison (Nhunda et al., 2026).

Table 8. Summary of findings — socio-ecological constraints.

3.4 Synthesis of results

Findings are organised into five hierarchical sub-findings, each concluding with a summary-of-findings table and figure placeholder consistent with the PRISMA 2020 reporting structure (Page et al., 2021a).

3.4.1 Indigenous biological resources

Autochthonous microbial, plant, and faunal resources from SSA mining regions form the substrate for most included sources. Acidithiobacillus ferrooxidans-dominated consortia isolated from the Zambian Copperbelt and the Witwatersrand pyrite horizons underpin bioleaching and AMD attenuation studies. Autocthonous Copperbelt filamentous fungi (Aspergillus niger, Penicillium chrysogenum) demonstrate effective Cu/Co biosorption. Native AMF consortia, notably at the Palabora Cu operation, drive phytostabilisation-assisted remediation (Table 9). Macrotermes bellicosus and M. falciger mediate entomoremediation of Au/As tailings via mound-soil translocation and gut-microbiome mediated attenuation (Nhunda et al., 2026; Tindwa and Singh, 2023).

Table 9. Summary of findings — indigenous biological resources.

3.5 Laboratory-reported efficacy across methodological families

Across the 75-source corpus, bioleaching, bioaugmentation, phytoextraction, and biosorption report the highest laboratory-scale efficacy (Table 10; Figure 4), with bioleaching of Cu-bearing sulfides reaching 65–88% recovery under controlled conditions, and sulfate-reducing-bacteria mediated AMD treatment reducing dissolved metals by 70–95% in batch systems (Nhunda et al., 2026.). Phytoextraction studies typically report 30–60% above-baseline shoot concentrations of Cu, Co, Zn, and Ni over single growing seasons, whilst biosorption studies report Cu, Pb, and Cd loadings of 40–220 mg/g dry biomass depending on biomass pretreatment and pH regime (Nhunda et al., 2026.; Tindwa and Singh, 2023).

3.6 Laboratory-to-field translation evidence and gaps

Field-scale evidence remains disproportionately sparse: ≈ 70% of sources report bench-scale work, ≈ 22% pilot, and only ≈ 8% true field-scale (Marrufo et al., 2025; Nhunda et al., 2026.). Where pilot and field data exist, efficacy is attenuated, with bioleaching Cu recovery typically falling from 65–88% (lab) to 40–65% (pilot) and 25–55% (field), and phytoextraction Cu/Co uptake diminishing by 30–50% under conditions owing to seasonal water stress and lithological heterogeneity (Table 10; Figure 5). Evaluated field-scale exemplars include the Rhodes BioSURE Process at the Rustenburg PGMs operation, Acidithiobacillus ferrooxidans-based Cu bioleaching pilot trials along the Copperbelt, Palabora AMF-assisted Cu phytostabilisation trials, and the West African Macrotermes bellicosus entomoremediation studies (Nhunda et al., 2026.; Tindwa and Singh, 2023).

Table 10. Summary of findings — laboratory-reported efficacy across methodological families.

3.7 Socio-ecological constraints mediating translation in SSA

Stacked socio-ecological constraints shape translation of laboratory bioremediation efficacy to field deployment in SSA: artisanal and small-scale mining governance fragmentation, francophone/lusophone indexing deficits, land-tenure insecurity that precludes multi-season phytoremediation commitments, water-availability variability driven by climate seasonality, and weak regulatory enforcement of mine-closure and tailings-management statutes (Adranyi et al., 2024; Marrufo et al., 2025; Nhunda et al., 2026; Tindwa and Singh, 2023). The ASM-context constraint dominates the Ghana, Tanzania, and DRC fringes of the corpus and reframes biological remediation from a purely technological fix to a socio-technical intervention (Adranyi et al., 2024).

3.8 Cost, beneficiation and recovery considerations

Reported bioleaching CAPEX is typically 30–50% lower than conventional pyro- or hydrometallurgical routes for low-grade Cu/Co ore, with OPEX dominated by acid/energy inputs and microbial-medium replenishment; NPV/IRR figures for SSA case studies span break-even to moderately positive returns contingent on ore grade and acid-consumption regimes (Nhunda et al., 2026). Biosorption offers a low-throughput, low-CAPEX secondary polishing step for AMD-laden effluents (Table 11; Figure 6) (Kara et al., 2023), with reported effluent reductions of 60–90% mg/L Cu/Pb/Cd against regulatory discharge limits. Phytoremediation endpoints remain diffuse, with recovery yields an order of magnitude below conventional routes but uniquely suited to diffuse, low-concentration, large-area contexts (Dusengemungu et al., 2022a; Marrufo et al., 2025; Nhunda et al., 2026.; Tindwa and Singh, 2023).

Table 11. Summary of findings — cost, beneficiation, and recovery.

4 Discussion

4.1 Summary of evidence

4.1.1 Indigenous biological resources as the substrate for SSA bioremediation

The synthesis confirms that Sub-Saharan Africa possesses a regionally distinctive and biologically diverse portfolio of candidate biological resources for the bioremediation of mining-derived contamination. Autochthonous microbial consortia from the Zambian Copperbelt include metal-resilient genera such as Aspergillus, Penicillium, Fusarium, Oidiodendron, Toxicladosporium, Septoglomus, and Neocosmospora. These taxa demonstrate tolerance to Cu and Co, which dominate the regional tailings geochemistry. The consortia are dominated by Basidiomycota (64.59%), Ascomycota (21.30%), Glomeromycota (4.53%), and trace Rozellomycota (0.0275%) (Dusengemungu et al., 2023) Filamentous fungi from the same metallogenic belt have been characterised as exceptionally capable, cost-effective biosorbents for Cu and Co recovery, with autochthonous isolates exhibiting biomass loadings and surface-chemistry traits that justify their candidacy for industrial-scale biosorption; the regional Cu/Co reserves are explicitly invoked as the rationale for deepening indigenous mycobiota characterisation (Dusengemungu et al., 2023). Pooled fungal isolates from Copperbelt legacy sites, including the Nkana Slag Dumpsite (“Black Mountain”) and Uchi Tailings Dam 26, have been deployed as blended inocula and shown to biomineralise Cu and Co in contaminated substrates (Dusengemungu, et al., 2022a; Mubemba, et al., 2022). A systematic review of bacterial bioremediation across SSA has documented 26 types of metal waste, with Bacillus and Pseudomonas genera prevalent across multiple metals and demonstrating high removal efficiencies under laboratory conditions (Uwimbabazi and Ramasamy, 2026).

Native plant resources are equally rich. The Copper–Cobalt Belt hosts more than 600 metallophytes and over 30 confirmed hyperaccumulators, with documented Cu shoot concentrations up to 13,700 mg kg^–1 in Aeollanthus subacaulis var. linearis, 12,300 mg kg^–1 in Ipomoea alpina, 9,322 mg kg^–1 in Crepidorhopalon perennis and 9,222 mg kg^–1 in Haumaniastrum katangense, alongside Co sequestration up to 10,232 mg kg^–1 and Ni 1,000–10,000 mg kg^–1 in Congolese and Zimbabwean Pearsonia metallifera and Dicoma niccolifera populations (Mubemba, et al., 2022). The geochemistry of terrestrial plants in the Central African Copperbelt has been systematically reviewed, confirming the successful application of hyperaccumulator species for phytoremediation and providing a basis for adopting phytogeochemistry in mineral exploration (Mukube et al., 2024). Recent investigations have also identified Persicaria puncata as a Co hyperaccumulator (up to 1,060 μg g^–1 in leaves) from the Copperbelt (Faucon et al., 2023). Excluder-type grasses and sedges co-occurring on the same substrates have been flagged as candidate phytostabilisers whose root-zone retention of metals limits offsite translocation. In Southern Africa, the Ni/Co hyperaccumulator Berkheya coddii has emerged as an agromining candidate with attainable metal yields of 77 kg ha^–1 Ni and 16.5 kg ha^–1 Co under optimum agronomic management (Rue et al., 2023). Recent pot experiments have further elucidated the interactions between Ni and Co for uptake and translocation in B. coddii, demonstrating that Co in the substrate limits Ni uptake (Rue et al., 2023). Continental-scale catalogues extend this inventory to 146 candidate phytoremediators, including 78 species native to Africa, supporting multi-element remediation of Ag, As, Au, Cd, Cu, Hg, Mn, Ni, Pb and Pt (Malunguja and Paschal, 2024). A comprehensive review focused on Africa has confirmed these findings, identifying 146 phytoremediators including 78 native African species across 57 herbs, 8 shrubs, and 13 trees (Anonymous, 2024a).

4.1.2 Laboratory-reported efficacy and the attenuation of bench-scale performance under field conditions

Laboratory studies report consistently high performance across the principal methodological families. Bioleaching of Cu and Co from primary sulfide and laterite sources has been characterised in comprehensive meta-analytic reviews distinguishing ore-grade from tailings-grade feedstocks and the microbial, geochemical and reactor parameters that govern recovery (Saim and Darteh, 2023). A comprehensive review on cobalt bioleaching from primary and tailings sources has examined the potential of different microorganisms to solubilise Co from sulfide and laterite deposits, providing an in-depth meta-analysis of research findings to enable inter-study comparisons (Saim and Darteh, 2023). Similarly, bioleaching of metal-bearing wastes and by-products has been reviewed, describing current operations and approaches at various scales and summarising the influence of operational parameters (Anonymous, 2023b; Kara et al., 2023; Malunguja and Paschal, 2024; Saim and Darteh, 2023). Stirred-tank and continuous-pilot bioleaching of pyrite-rich sulfidic tailings using BRGM-KCC consortia has achieved cobalt dissolution above 90%, sufficient to underpin a profitable hypothetical industrial operation (Anonymous, 2023b; Kara et al., 2023). A handful of bioleaching operations have reached industrial scale, including a stirred-tank cobalt-recovery process applied to a pyrite concentrate stockpiled at the former Kilembe copper mine in Kasese, Uganda — one of the few documented commercial-scale exemplars worldwide and an instructive comparator for the SSA context (Anonymous, 2023f; Roberto and Schippers, 2022). Sulfate-reducing-bacteria bioaugmentation in batch systems removes 70–95% of dissolved metals from AMD-impacted waters; recent studies confirm that organic substrate type strongly influences SRB performance, with sugarcane bagasse and shrimp shell supporting the highest AMD-treating capacity and sulfate removal efficiencies of 95.6% and 86.0%, respectively (Li et al., 2022); biosorption with fungal and agricultural-waste biomass removes 60–90% of Cu, Pb and Cd from effluent streams; and chelator-assisted phytoextraction supports 30–60% above-baseline shoot uptake of Cu, Co, Zn and Ni over single growing seasons (Anonymous, 2023e; Jeyakumar et al., 2022; Zulkernain et al., 2023).

Translation of these bench-scale efficacies into sustained field performance is, however, attenuated by a well-documented constellation of factors. Field-realistic influences—including contaminant concentration, fertiliser and chelator scheduling, planting density, soil texture and pH, salinity, and biosafety considerations—frequently erode bench-to-field transfer (Mohamed et al., 2025). In SSA-relevant North African case studies, metal accumulation in plant shoots often exceeded conventional thresholds and transfer factors were measured. However, no candidate species met Kabata-Pendias hyperaccumulator criteria. This leaves promising bench performance without formal field-deployable classification (Zulkernain et al., 2023). The Moroccan evidence base consolidates this pattern: despite progress in identifying candidate species and rhizosphere-assisted strategies, “all of the studied options remain in the testing stage” and the contamination issues they are intended to address continue to persist (Zulkernain et al., 2023). At the African regional scale, the translation gap has accordingly been characterised as a structural feature of the bioremediation evidence base rather than as a transient technical hurdle (Mohamed et al., 2025). A systematic review of bacterial bioremediation across SSA has similarly identified a restricted geographical scope and limited field-based application, which may limit understanding of field application of bacterial bioremediation (Uwimbabazi and Ramasamy, 2026). The review of phytoremediation in Africa also notes that the identification of suitable plants requires extensive laboratory testing, hindering practical application in developing African countries (Anonymous, 2024b).

4.1.3 Socio-ecological constraints and cost, beneficiation and recovery considerations in the SSA context

The persistent attenuation of laboratory performance in SSA is inseparable from the socio-ecological context in which bioremediation must operate. Artisanal and small-scale mining generates a disproportionately large share of the regional contamination burden, and mining-derived metals are tightly coupled to livelihoods, food security and water-supply systems (Marrufo et al., 2025). Within this labour landscape, 50–60% of the African artisanal and small-scale gold mining workforce is female, compared with approximately 10% globally. Women disproportionately undertake ore processing and amalgam burning. Documented sex-, age-, and PPE-mediated variation in mercury exposure exists, with limited baseline toxicity knowledge even amongst primary caregivers (Ondayo et al., 2023). A systematic review of environmental and health risks posed to children by artisanal gold mining estimated that 5 million children work in ASGM worldwide, with hazards poorly characterised and underreported (Allan-Blitz et al., 2022). Scoping evidence from 176 ASM health studies across 38 countries confirms that vulnerable groups — women of reproductive age and children — receive disproportionately little research attention despite carrying the greatest toxic-exposure burden, with infectious-disease and mental-health dimensions remaining underexplored (Anonymous, 2024a). A review of mercury pollution from ASGM in Sub-Saharan Africa has established that the sector contributes up to 20% of global gold production and uses 205–496 tonnes/yr of Hg, with the highest levels of mercury pollution in mining soils observed in West Africa (Anonymous, 2024a). These intersectional vulnerabilities are refracted through state-change dynamics in ASM landscapes: in Ghana, AGM has been quantifiably linked to land-cover change (≈ 1.13%), topography flattening and biota loss, with cascading consequences for ecosystem services and human welfare (Agyapong et al., 2026; Timsina et al., 2022). A review of tropical surface gold mining has confirmed that surface gold mining severely degrades landscapes, causing deforestation, soil erosion and displacement, and toxic contamination, with restoration strategies requiring specific attention to the unique ecological conditions of former gold mines (Timsina et al., 2022). Bioremediation is therefore increasingly framed as a socio-technical intervention alongside its technological rationale (Ripanda et al., 2025).

Cost, beneficiation and recovery considerations sharpen this framing. Biohydrometallurgical routes to metal recovery operate with markedly lower installation and operating costs than pyrometallurgical and hydrometallurgical alternatives, and life-cycle assessment paired with techno-economic analysis has been recommended as the methodological standard for quantifying environmental and economic profiles before any fielded deployment (Kara et al., 2023). Mineral-recovery economics can underwrite remediation: a conceptual model for SSA specifically proposes that revenue from phytoextracted metals can subsidise multi-hectare phytoremediation operations and provide livelihood support to mining-affected communities, provided socio-economic barriers are addressed (Timsina et al., 2022). South African agromining with B. coddii offers one of the few quantified metal-crop yield datasets available for SSA (Rue et al., 2023). For bioleaching of Congolese Copperbelt polymetallic Cu/Co/Ni concentrates, technical feasibility has been demonstrated at laboratory scale, but industrial-scale mobilisation is constrained by the large upfront capital required for new plants, by the inherently slow kinetics of bacterial leaching in columns and heaps, and by the delayed acid generation that offsets OPEX only over time, whilst economic analyses for tailings reprocessing are still rare across the region (Roberto and Schippers, 2022). Financial assurance programs have been evaluated as pivotal in promoting responsible mining practices and ensuring sustainable mine closure in South Africa, Zambia, Ghana, and Kenya, with key institutional design variables identified as almost interrelated across these jurisdictions (Anonymous, 2023a).

4.2 Limitations of the evidence and review

Several limitations circumscribe the inferences that can be drawn from this synthesis. First, the corpus is structurally laboratory-skewed: most included sources report batch, microcosm or greenhouse pot trials under tightly regulated parameters that under-represent the heterogeneity, multi-metal co-contamination, climate seasonality and biological interactions characteristic of SSA field sites (Mohamed et al., 2025; Otunola and Mhangara, 2025). This skew is not unique to the region but is amplified locally by infrastructural constraints that limit the establishment of long-term field trials (Chavula and Kayusi, 2025; Nhunda et al., 2026). Second, geographic and linguistic under-representation is pronounced: francophone and lusophone SSA jurisdictions—including the Democratic Republic of Congo’s Katanga Copperbelt and Angola’s emerging mining regions—remain substantially absent from mainstream Anglophone bibliographic indexing, biasing the synthesis towards the Copperbelt and the Witwatersrand and limiting external validity across the region’s non-Anglophone mining geographies (Agyapong et al., 2026; Marrufo et al., 2025; Mulenga et al., 2024; Peprah, 2024). This linguistic filtration is a recognised structural bias in environmental evidence synthesis from SSA that systematically depresses the visibility of central and southern African research outputs (Tindwa and Singh, 2023). The systematic review of bacterial bioremediation across SSA similarly identified a restricted geographical scope as a key limitation (Uwimbabazi and Ramasamy, 2026).

Third, the adapted-CASP risk-of-bias assessment identified C5 (baseline characterisation and translational transparency) as the criterion of greatest concern. Parallel geochemical baselines—pH, Eh, particle-size distribution, total versus bioavailable metal speciation—and scale-of-execution controls were frequently absent. This undermines cross-source comparability and laboratory-to-field extrapolation (Härtung et al., 2025; Page et al., 2021a; Whaley et al., 2022).The reporting-quality deficits flagged here are consistent with broader critiques of environmental remediation studies in low- and middle-income contexts. Fourth, microbial protocols are non-standardised across the corpus, with limited replication of media composition, temperature regimes and inoculum preparation, complicating pooled interpretation (2023; Jeyakumar et al., 2022; Saim and Darteh, 2023). The absence of harmonised reference strains and positive controls further reduces the capacity to distinguish method-attributable effects from isolate-specific idiosyncrasies (Gwanama et al., 2022).

Fifth, methodological heterogeneity—across analytes, units, time horizons, baseline conditions and reporting formats—precluded quantitative meta-analytic synthesis, restricting the review to descriptive-quantitative handling, and may itself discourage the field-trial programmes that commercial-scale phytomining and agromining require (Kara et al., 2023; Tezyapar Kara et al., 2022). This heterogeneity has been identified as a persistent barrier to evidence aggregation in tropical bioremediation research more broadly (Kara et al., 2023). Sixth, even where phytoremediation achieves field-measured uptake, food-chain and biomass-disposal pathways continue to limit deployment and require explicit risk management (Kara et al., 2023). The potential for metal remobilisation through composting, incineration or landfilling of harvested tissues remains inadequately addressed in the SSA literature (Otunola and Mhangara, 2025).

Finally, several review-level limitations merit explicit acknowledgement: the adapted-CASP instrument prioritises internal validity over external ecological validity (Frampton et al., 2022); the literature search, whilst structured, cannot preclude missed grey-literature or non-indexed francophone and lusophone work (Qadir et al., 2024; Frampton et al., 2022); and dual-reviewer consensus, although arbitrated by a third reviewer, remains subject to unconscious disciplinary bias (Shaheen et al., 2023; Frampton et al., 2022).

4.3 Pathways to practice, policy, and future research

Practice and technology deployment. Technology-transfer pipelines for SSA bioremediation should be organised as tiered lab-to-pilot-to-field scale-up protocols with mandatory baseline geochemical characterisation at each transition, including pH, Eh, particle-size distribution, total and bioavailable metal speciation, and explicit documentation of media, temperature, inoculum preparation and replicated controls (Curiel Alegre et al., 2024; Dusengemungu, et al., 2022a; Mubemba, et al., 2022; Nhunda et al., 2026). Locally sourced electron donors, agricultural-waste-derived amendments (biochar, compost, manures) and indigenous microbial and AMF consortia should be prioritised over external inocula to reduce biosecurity risks—including the documented horizontal gene-transfer concerns associated with introduced metal-resistant strains—and to align with regional cost structures (Ayilara and Babalola, 2023; Dusengemungu et al., 2023; Zulkernain et al., 2023). For phytoremediation and agromining specifically, agronomic protocols (plant density, cropping cycles, harvesting regimens) need to be co-designed with ASM communities and matched to customary land-use patterns so that contamination remediation and livelihood protection are pursued jointly (Kara et al., 2023). Recent evidence from the Copperbelt region of Africa confirms that phytoremediation programs have been initiated in some tailings, with Leucaena leucocephala identified as a Cu/Co excluder species that accumulates bioavailable metals from the rhizosphere in its roots (Lubobo et al., 2024).

Policy and governance. Bioremediation cannot be operationalised at scale without effective ASM formalisation, mine-closure compliance and tailings-management enforcement (Adranyi et al., 2024). Across SSA, policy coherence between mining, agriculture, water, land, environment and development portfolios is reportedly partial rather than full, with inter-sectoral coordination emerging as the binding constraint rather than entrenched sectoral resistance (Adranyi et al., 2024). Tailings-management and remediation-technology requirements should be embedded in mine-closure plans and backed by enforceable closure bonds, with explicit fiscal and technical support for ASM formalisation pathways (Kara et al., 2023); West African evidence indicates that formalisation without operator support has, in some jurisdictions, delivered state control without proximate livelihood benefit (Adranyi et al., 2024). Financial assurance programs have been evaluated across South Africa, Zambia, Ghana, and Kenya, with problems associated with executing financial guarantees identified as including governments’ deficiency of understanding of the practice of financial assurance (Anonymous, 2023b). Indexing and Francophone/Lusophone representation in regional evidence syntheses must be strengthened: francophone and lusophone bibliographic coverage needs active investment, including translation support and inclusion of regional theses and grey literature (Agyapong et al., 2026; Marrufo et al., 2025; Mulenga et al., 2024). Mining-derived metal pollution more broadly continues to pose documented ecological and public-health risks to SSA communities, underscoring the urgency of policy alignment (Pathé et al., 2025; Ripanda et al., 2025; Tindwa and Singh, 2023). Local responses to mine closure in South Africa have been investigated, linking dependencies and social disruption, with implications for closure cost frameworks and policy development (Marais et al., 2024).

4.3.1 Future research priorities

Six priority research streams are warranted. (i) Standardised reporting protocols for SSA bioremediation, building on the PRISMA 2020 and ROSES frameworks, are needed to harmonise outcome metrics, baseline descriptors and adverse-event reporting across the corpus (Page et al., 2021a). (ii) Longitudinal, multi-season field monitoring of pilot-scale interventions is required to capture climate-seasonality effects on biological endpoints—including drought stress, altered precipitation, and elevated CO₂ dynamics that affect metal bioavailability, plant uptake, microbial activity and rhizosphere processes (Mohamed et al., 2025). Remote sensing and GIS integration has been demonstrated as a scalable approach for monitoring mining-related vegetation change, with Landsat time-series segmentation achieving 92.6% accuracy for ASM detection (Ouédraogo et al., 2024; Fonseca et al., 2024). Assessment of land cover degradation due to mining activities using remote sensing and digital photogrammetry has confirmed significant loss of vegetation cover and increase in bare soil (Blanche et al., 2024). (iii) Cost-benefit and beneficiation analyses, integrating life-cycle assessment and techno-economic analysis as standard, are prerequisites for justifying deployed remediation investments and for evaluating recovery economics alongside environmental performance (Kara et al., 2023; Tezyapar Kara et al., 2022). (iv) Sex-, age- and vulnerability-differentiated impact-assessment protocols—extending well beyond the descriptive exposure baselines currently available—are needed to ensure that remediation gains translate into measurable health benefits for women, children and other exposed groups, who remain under-represented in the ASM health-studies literature (Ondayo et al., 2023). The systematic review of environmental and health risks to children in ASGM has highlighted profound underreporting of the prevalence and consequences of hazards amongst children (Allan-Blitz et al., 2022). (v) Climate-resilient remediation design must be integrated from project inception given the documented vulnerability of metal bioavailability and phytoremediation efficacy to climate stressors (Mohamed et al., 2025; Patel and Srivastava, 2025). (vi) Integration of remote sensing and GIS with biological endpoints—using Sentinel-2 multi-temporal spectral indices for phytostabilised versus natural vegetation, airborne hyperspectral imagery with optimised vegetation indices for canopy-level metal prediction (r ≥ 0.84 and RPD ≥ 2.06 reported for Cu and Zn), and platform-agnostic spectrotemporal indices linked to MODIS and Landsat archives—offers a scalable, non-invasive framework for monitoring SSA mining regions over the multi-year horizons that biological remediation demands (Galodha et al., 2025). Enhanced detection of ASM using spectral and textural segmentation of Landsat time series has demonstrated the feasibility of large-area monitoring from space with free and open-source GEE and operational satellites (Fonseca et al., 2024).

5 Conclusion

This PRISMA-2020-aligned review synthesises 75 sources on the bioremediation of mining waste and tailings in SSA (January 2022 to mid-2026) and arranges the evidence along four principal threads.

First, SSA possesses a regionally distinctive autochthonous biological resource base. Copperbelt mycobiota—dominated by Basidiomycota (64.6%) and Ascomycota (21.3%)—display documented Cu/Co tolerance. Central-African hyperaccumulators such as Aeollanthus subacaulis var. linearis (shoot Cu up to ~13,700 mg kg^–1), Ipomoea alpina, Crepidorhopalon perennis, and Haumaniastrum katangense provide high-biomass phytoextraction candidates (Marques and Rodrigues, 2024; Kara et al., 2023). Entomoremediation taxa (Macrotermes bellicosus, M. falciger) and AMF consortia underpin Palabora-type phytostabilisation. Second, laboratory efficacy across the eight methodological families is consistently high — bioleaching Cu/Co recovery 65–88% (Kara et al., 2023; Roberto and Schippers, 2022; Saim and Darteh, 2023) and SRB-mediated dissolved-metal removal 70–95% (Jeyakumar et al., 2022; Anonymous, 2023c; Zulkernain et al., 2023). Third, bench gains are persistently attenuated under field conditions (factors ≈0.20–0.55), with the attenuation driven by ASM governance fragmentation, francophone and lusophone indexing deficits, land-tenure insecurity, weak regulatory enforcement of mine-closure and tailings-management statutes, and subsistence-economy coupling of contamination exposure (Marrufo et al., 2025; Anonymous, 2023d; Ripanda et al., 2025). Fourth, the cost–beneficiation–recovery envelope positions bioremediation as a socio-technical rather than purely technological intervention: biohydrometallurgical routes offer CAPEX advantages and locally available substrates reduce OPEX (Anonymous, 2023e; Curiel Alegre et al., 2024), but no techno-economic case alone resolves the documented lab-to-field attenuation.

5.1 Implications

For practice, translation pipelines should be organised as tiered lab-to-pilot-to-field protocols with mandatory baseline geochemical characterisation at each transition. Locally sourced electron donors and indigenous microbial and AMF consortia should be prioritised. Joint livelihood–remediation design, co-developed with ASM-affected communities and embedding recovered-metal revenue in benefit-sharing frameworks, is a precondition for uptake (Calderon et al., 2021; Mohamed et al., 2025; Anonymous, 2024a; Rue et al., 2023). For policy, ASM formalisation supported by enforceable closure bonds, tailings-management requirements embedded in mine-closure plans, and active investment in francophone and lusophone bibliographic coverage are the binding preconditions for any scaled deployment (Khan et al., 2026; Adranyi et al., 2024; Marrufo et al., 2025). For research, the field’s most urgent priorities are: SSA-specific reporting protocols building on PRISMA 2020; longitudinal, multi-season field monitoring under climate variability; life-cycle–techno-economic integration as the analytical standard; sex-, age-, and vulnerability-differentiated impact assessment responsive to ASM’s female-labour share and the disproportionate toxic-exposure burden on women and children (Ondayo et al., 2023); climate-resilient remediation design from inception (Mohamed et al., 2025; Patel and Srivastava, 2025); and integration of Sentinel-2 multi-temporal spectral indices, airborne hyperspectral imagery, and platform-agnostic spectrotemporal indices with biological endpoints for non-invasive multi-year monitoring (Moyo et al., 2023; Galodha et al., 2025).

Under mounting ASM-driven pollution, climate-stress intensification, and a widening ecological and public-health burden, the case for prioritising indigenous biological capital, locally rooted deployment architectures, and the cross-sectoral policy alignments without which the laboratory-to-field gap will persist is decisive.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Mohammad Bagher Hassanpouraghdam, University of Maragheh, Iran

Reviewed by: Khuram Shehzad Khan, China Agricultural University, China

Francisco J. Bedoya Rodríguez, University of Santiago de Cali, Colombia