Characterization and enhancement of household solid waste composting using effective microorganisms

Household organic solid waste is a growing environmental burden in Zanzibar. This study evaluated effective microorganism (EM) inoculation as a means of accelerating household organic waste composting and improving final compost quality. Composting was conducted over 60 days in triplicate bins of three working volumes (6, 9, and 21 L) filled to approximately 60% of capacity, comparing two treatments: EM-inoculated and an uninoculated control. Activated EM was applied as a liquid inoculum at a fixed rate scaled to feedstock load (10 mL of activated EM working solution per kg wet feedstock) and mixed homogeneously with the feedstock. Temperature, moisture, pH, and electrical conductivity (EC) were monitored throughout; chemical indices (organic matter, C/N ratio, N–P–K, humic substances, heavy metals) and biological indicators (germination index, microbial counts, Shannon diversity, qPCR/sequencing of dominant taxa) were measured in mature compost. EM-inoculated bins reached 65 °C within 3 days vs. 55 °C in controls and sustained thermophilic conditions (≥55 °C) for 10 vs. 6days (F = 8.23, p = 0.01). Moisture remained within 50%−60% in EM treatments (p = 0.03). EM compost showed faster organic matter loss (28% vs. 35% at day 60; p = 0.02), higher nitrogen retention (2.1% vs. 1.7%; p = 0.01), improved phosphorus and potassium content (+15% and +18%; p < 0.05), a lower final C/N ratio (18 vs. 22; p = 0.02), earlier EC stabilization (3.5 vs. 4.0 dS/m; p = 0.04), greater humification (+35% vs. +20%; p = 0.03), higher microbial diversity (Shannon 3.8 vs. 3.1; p = 0.02), and superior maturity (germination index 85% vs. 70%; p = 0.01). Heavy metals remained below permissible limits in all treatments (p > 0.1). These findings indicate that EM inoculation accelerates composting, yields a nutrient-dense and phytotoxicity-safe product, and can support decentralized waste management in resource-constrained communities.

Municipal solid waste generation is increasing worldwide as populations grow and urbanize, placing strain on collection, treatment, and disposal systems (Barma et al., 2022). In many low- and middle-income countries the problem is compounded by inadequate infrastructure, open dumping, uncontrolled burning, and weak policy enforcement (Almokmesh et al., 2024; Bundhoo, 2018). Organic materials food scraps, yard trimmings, and other biodegradable household residues typically constitute more than half of domestic waste streams, yet recycling and composting rates remain low (Manea et al., 2024).

Zanzibar illustrates these challenges. Limited waste collection coverage, poor disposal practices, and low community participation leave organic waste largely unmanaged, contributing to environmental contamination and public-health risks (Mukama et al., 2016). Composting offers a practical route for diverting organics from landfills while generating nutrient-rich soil amendments that support agriculture and reduce greenhouse gas emissions (Larney et al., 2006; Manea et al., 2024). The approach is particularly attractive for decentralized settings because it requires modest capital investment and can be scaled to household or community level (Katranci et al., 2026).

Effective microorganisms (EM) are mixed-culture inoculants containing lactic acid bacteria, photosynthetic bacteria, yeasts, and actinomycetes that have been reported to accelerate organic matter decomposition, suppress pathogens, and reduce odors during composting (Weissert et al., 2025; Greff et al., 2021). Laboratory and field trials on substrates ranging from rice straw to mushroom waste have shown that EM application can shorten composting duration, raise peak temperatures, and improve compost nutrient content and maturity (Ansari et al., 2022; Jusoh et al., 2013; Raimi et al., 2024). However, much of this work was conducted on agricultural or industrial feedstocks rather than mixed household organic waste, and few studies have combined physical, chemical, and molecular-biological indicators within a single experimental design. The novelty of this study lies in its integration of physical, chemical, and molecular-biological indicators (qPCR and amplicon sequencing) within a single controlled composting experiment on mixed household organic waste under Zanzibari conditions an integrated, multi-indicator evaluation of EM inoculation not previously reported for domestic organic waste in a Sub-Saharan island setting.

The present study therefore aimed to (1) characterize the composting process of household organic solid waste inoculated with EM relative to uninoculated controls, and (2) quantify the effects of EM inoculation on process performance (temperature, moisture, pH, EC), chemical composition (organic matter, C/N, N–P–K, humic substances, heavy metals), microbial community dynamics, and compost maturity (germination index). The results are intended to inform low-cost, decentralized waste management strategies applicable in Zanzibar and comparable resource-constrained settings.

Experimental design

Composting experiments were conducted in controlled bins of three working volumes (6, 9, and 21 L). Each bin was loaded with source-separated household organic waste (predominantly food waste and yard trimmings) to approximately 60% of its capacity, leaving sufficient void space for aeration; this corresponded to a mean fresh mass of 2.0 ± 0.2 kg in the 6 L bins, 3.5 ± 0.3 kg in the 9 L bins, and 10.0 ± 0.8 kg in the 21 L bins. Two treatments were compared: (1) EM-inoculated bins and (2) uninoculated control bins receiving no microbial additive. Activated EM was applied as a liquid inoculum at a single fixed rate scaled to feedstock load (10 mL of activated EM working solution per kg wet feedstock) and mixed homogeneously with the feedstock; the previously reported 20–100 mL range refers to the absolute inoculum volumes delivered to the smallest and largest bins, respectively, under this fixed rate rather than to different doses per treatment. Each treatment was replicated three times. Each treatment comprised one bin at each of the three working volumes (6, 9, and 21 L); the three replicates per treatment therefore corresponded to the three different bin volumes rather than to independent bins of a single volume. A total of six bins was used (three EM-inoculated and three uninoculated controls). Bin volume was treated as a blocking factor in the analyses, with each volume contributing one EM-inoculated and one control bin; because the volumes were not independently replicated within each treatment, the bin-volume effect cannot be fully separated from the treatment effect, and this constraint is acknowledged as a design limitation in Section 4.5. Composting duration was 60 days, with conditions maintained to promote thermophilic development: target moisture of approximately 60%, periodic manual aeration, and daily temperature monitoring (Raimi et al., 2024; Lew et al., 2021). The vessels were locally sourced, cylindrical plastic containers with nominal capacities of 6, 9, and 21 L. Following the bin configuration described by (Raimi et al. 2024), aeration holes approximately 0.7 mm in diameter were drilled around the sidewalls at about 5 cm spacing. A mesh support was positioned approximately 3 cm above the base to retain the feedstock while leaving a lower void for leachate collection. The bins had loose-fitting lids, no external insulation, and were maintained under the same sheltered ambient conditions. Because the containers were recorded by nominal capacity, their exact internal height and diameter and the total number of aeration holes were not retained in the study record and cannot be reconstructed reliably (Supplementary Table S1). At loading, the feedstock and inoculum were mixed thoroughly by hand. During composting, all bins were turned manually with a hand trowel under the same procedure and schedule; however, the precise turning interval and duration were not recorded. Moisture was monitored gravimetrically, and dechlorinated water was added when moisture fell below approximately 55% to restore the target of about 60%; the dates and volumes of these additions were not logged per bin. These record-keeping limitations are reported explicitly rather than replaced with retrospectively estimated values.

Sample collection and preparation

Household solid waste was sampled using a stratified random design across households of different socioeconomic strata to capture variability in waste generation and composition (Bolaane and Ali, 2004; Monavari et al., 2011). On a wet-weight basis, the pooled organic feedstock comprised approximately 55% food and kitchen waste, 25% vegetable and fruit residues, 15% yard and garden trimmings, and 5% other biodegradable organics Figure 1. Collected waste was sorted into organic and inorganic fractions; the organic fraction was shredded to a particle size below 20 mm, homogenized, and loaded into composting bins. Segregation bins were installed at participating households to facilitate source separation and improve sample quality (Requena-Sanchez et al., 2022).

EM inoculum was prepared by activating a commercial EM stock culture in a molasses-water medium (1:1:20 v/v) under anaerobic conditions for 7 days at ambient temperature, following standard preparation protocols. The activated EM, containing lactic acid bacteria, photosynthetic bacteria, yeasts, and actinomycetes, was further diluted in dechlorinated water to a working concentration of 1:100 (v/v), corresponding to approximately 1% (v/v) activated EM in the applied solution and applied at the time of bin loading at the fixed rate specified in Section 2.1 (Jusoh et al., 2013; Raimi et al., 2024). The stock culture was a commercial EM inoculant obtained from a local agricultural supplier in Zanzibar. The product name, manufacturer, batch number, and exact ambient activation temperature were not retained in the available laboratory record. The viable concentration of the activated stock was not measured before application; consequently, a final CFU/mL value cannot be reported. The recorded activation ratio, seven-day anaerobic incubation, working dilution, and application rate are therefore reproducible, while the unavailable product identity, activation temperature, and initial viable count are acknowledged as methodological limitations.

Analytical methods

Physical parameters were measured as follows. Temperature was recorded daily at three depths within each bin using calibrated thermocouples. Moisture content was determined gravimetrically by drying sub-samples at 105 °C to constant weight. pH and electrical conductivity (EC) were measured in 1:10 (w/v) aqueous extracts using calibrated meters.

Chemical analyses included organic matter content by loss on ignition at 550 °C, total nitrogen by Kjeldahl digestion, total phosphorus by spectrophotometry after acid digestion, and total potassium by flame photometry. The C/N ratio was calculated from organic carbon (OM/1.724) and total nitrogen values. Humic substance fractionation (humic and fulvic acids) followed the IHSS alkali-extraction protocol. Heavy metals (Pb, Cd, Zn) were determined by atomic absorption spectrometry after acid digestion (Gao and Xu, 2014).

Microbiological assessments combined culture-based and molecular methods. Total viable counts of thermophilic and mesophilic bacteria were determined by serial dilution and spread-plating on nutrient agar incubated at 55 °C and 30 °C respectively; colonies were enumerated after 48 h and expressed as CFU per g dry compost (n = 3). For molecular analysis, total genomic DNA was extracted from 0.25 g of compost using the DNeasy PowerSoil Pro Kit (QIAGEN, Hilden, Germany), and DNA quality and quantity were verified by NanoDrop spectrophotometry (Thermo Fisher Scientific) and 1% agarose gel electrophoresis. Bacterial community composition was characterized by 16S rRNA gene amplicon sequencing of the V3–V4 region using primers 341F (5′-CCTACGGGNGGCWGCAG-3′) and 805R (5′-GACTACHVGGGTATCTAATCC-3′) on the Illumina MiSeq platform (2 × 300 bp paired-end). Raw reads were processed with QIIME 2 (v. 2024.2) with the DADA2 plugin: reads were quality-filtered, denoised into amplicon sequence variants, and taxonomically classified against the SILVA 138 database. Shannon diversity indices and taxon relative abundances were computed from the resulting feature table rarefied to 20,000 reads per sample. EM-associated taxa (lactic acid bacteria, photosynthetic bacteria, yeasts, and actinomycetes) were quantified by quantitative PCR (qPCR) using the group-specific primers and thermocycling, with abundances expressed as gene copies per g dry compost relative to standard curves (amplification efficiency 95%−102%, R 2 ≥ 0.99). Compost maturity was assessed by the germination index (GI) using cress (Lepidium sativum) seed bioassays (Cangelosi and Meschke, 2014; Nemati et al., 2016).

Data analysis

All data were tested for normality (Shapiro–Wilk) and homogeneity of variances (Levene's test) prior to analysis. Differences between EM-inoculated and control treatments were evaluated using independent-samples t -tests for single time-point comparisons and repeated-measures ANOVA for time-series data (temperature, moisture). One-way ANOVA was applied to compare treatment means for chemical and biological parameters at maturity. Because only three replicates were available per treatment, the assumptions of the parametric tests were verified for each parameter before application; where normality or homoscedasticity was not satisfied, non-parametric alternatives (Mann–Whitney U or Kruskal–Wallis tests) were used as a safeguard; in the present dataset, however, all parameters satisfied the normality and homoscedasticity assumptions, so the parametric tests reported in Tables 1–4 were retained throughout, and because only three replicates were available per treatment these results should be interpreted with caution given the limited statistical power. To account for the three bin volumes, bin volume was included as a blocking factor in the analysis; because each volume was represented by only one bin per treatment, however, its effect could not be fully separated from that of treatment, and this confound is addressed in Section 4.5. The test used, the corresponding test statistic with degrees of freedom, and the exact p -value are reported for every parameter in Tables 1–4. Significance was set at α = 0.05. Analyses were performed with R (v. 4.4) (Abbasnasab Sardareh et al., 2021). It should be noted that normality and homogeneity of variance are independent assumptions; applying a test for equal variances therefore does not imply that the data departed from normality, and both diagnostics were run as standard practice. Levene's test (based on absolute deviations from the group median) was chosen in preference to Bartlett's test because, although Bartlett's test is more powerful when the data are exactly normal, it is highly sensitive to even minor departures from normality, whereas Levene's test is robust to such departures. Given the small replication (n = 3 per treatment), which limits the power of the Shapiro–Wilk test to detect non-normality, the more conservative Levene's test was retained to avoid over-reliance on a normality assumption that could not be established with confidence.

Physical characteristics during composting

Temperature profiles for the two treatments are presented in Figure 2. EM-inoculated compost reached a peak thermophilic temperature of 65 °C within 3 days, compared with 55 °C in controls. Temperatures above 55 °C were sustained for 10 days in EM bins vs. 6 days in controls; this difference was statistically significant (ANOVA, F = 8.23, p = 0.01). Both treatments returned to ambient levels by day 40, though the EM treatment remained consistently warmer during the first three weeks.

Moisture content in EM-treated compost stayed within the optimal range of 50%−60% throughout the process, with fluctuations limited to ±5%, whereas control compost ranged from 45 to 65% and occasionally dropped below 40% (t -test, p = 0.03). pH rose from an initial 6.5 to 8.2 during the thermophilic phase in EM compost and stabilized at 7.5 at maturity; controls stabilized at 7.0 (p < 0.05). EC increased from 1.2 to 3.5 dS/m in EM compost but leveled off earlier than in controls, where EC reached 4.0 dS/m before stabilizing (repeated-measures ANOVA, p = 0.04).

Chemical composition of mature compost

Organic matter content declined from an initial 65% (dry weight) to 28% in EM-inoculated compost by day 60, compared with 35% in controls (p = 0.02), confirming accelerated decomposition under EM treatment (Figure 3). The C/N ratio dropped from 30 to 18 in EM compost and to 22 in controls over the same period (p = 0.02), reflecting faster carbon mineralization and better nitrogen conservation Figure 4 show.

Total nitrogen averaged 2.1% dry weight in mature EM compost vs. 1.7% in controls (p = 0.01). Phosphorus and potassium concentrations were approximately 15% and 18% higher in EM treatments (P: 0.45% vs. 0.39%; K: 1.2% vs. 1.0%; both p < 0.05). All chemical parameters are reported as mean ± SD (n = 3) in Table 2. Heavy-metal concentrations (Pb, Cd, Zn) are reported as mean ± SD in Table 2 (Pb: 12.4 ± 1.8 vs. 13.1 ± 2.0 mg/kg; Cd: 0.42 ± 0.08 vs. 0.46 ± 0.09 mg/kg; Zn: 84 ± 11 vs. 88 ± 12 mg/kg). All values were below the limits set by the US EPA 40 CFR Part 503 Class A (exceptional-quality) limits for land-applied biosolids (Pb < 300, Cd < 39, Zn < 2,800 mg/kg dry weight), with no significant difference between treatments (p > 0.1), indicating that EM inoculation did not introduce additional contamination.

Microbial community dynamics

Thermophilic bacterial populations peaked at 1.2 × 10 8 CFU/g in EM compost on day 7, significantly higher than 7.5 × 10 7 CFU/g in controls (p = 0.03). Mesophilic populations increased during maturation in both treatments, but EM compost maintained roughly 20% higher microbial diversity throughout the process (Shannon index: 3.8 vs. 3.1 at maturity; p = 0.02).

Amplicon sequencing indicated that Firmicutes (notably Bacillus spp.) dominated the thermophilic phase, while Actinobacteria and Proteobacteria increased in relative abundance during cooling and maturation; the corresponding relative-abundance values and taxonomic composition Fungal communities were dominated by Ascomycota, with Cladosporium and Penicillium present at relative abundances of approximately 15% and 10% of fungal reads, respectively. Quantitative PCR indicated that EM-associated taxa (lactic acid bacteria, photosynthetic bacteria, yeasts, and actinomycetes) were more abundant in EM bins than in controls—by 1.3- to 1.5-fold in gene-copy abundance (lactic acid bacteria 8.4 ± 0.9 vs. 6.1 ± 0.7 × 10 7, photosynthetic bacteria 3.2 ± 0.5 vs. 2.4 ± 0.4 × 10 6, yeasts 1.5 ± 0.3 vs. 1.1 ± 0.2 × 10 6, and actinomycetes 9.7 ± 1.1 vs. 7.8 ± 0.9 × 10 6 copies per g dry compost; independent-samples t -tests, p = 0.01–0.04; a 10%−25% higher relative abundance; p < 0.05)—consistent with establishment of the inoculated consortium.

Compost maturity and stability

The germination index (GI) of EM-inoculated compost reached 85% after 45 days, compared with 70% in controls (p = 0.01). A GI above 80% is generally considered indicative of mature, non-phytotoxic compost (De Moraes Cunha Goncalves et al., 2020). Humic substance content increased by 35% relative to initial levels in EM compost, significantly exceeding the 20% increase in controls (p = 0.03). Together with the lower C/N ratio and higher GI, these results indicate that EM treatment produced a more mature and stable end product Figure 5.

Temperature, moisture, and pH dynamics

The rapid attainment of thermophilic temperatures in EM-inoculated bins (65 °C within 3 days) and the extended duration above 55 °C are consistent with reports that EM inoculation stimulates early microbial metabolic activity, generating greater heat output from labile substrate decomposition (Mckinley and Vestal, 1985; Lai et al., 2025). Thermophilic temperatures above 55 °C for several consecutive days are important for pathogen destruction under international composting standards, so the 10-day exceedance in EM bins represents a practical advantage for sanitation. It should be noted, however, that the association between the higher temperatures and microbial abundance was not evaluated by formal correlation analysis, and that composting temperature is additionally governed by bin volume, heat-retention capacity, substrate composition, moisture, and aeration. The elevated temperatures in the EM bins therefore cannot be attributed to greater EM abundance alone; the co-occurrence of higher temperature and higher EM-associated gene-copy numbers is presented as a hypothesis to be confirmed by future work relating time-resolved microbial counts to temperature within each bin volume.

Moisture stability in the 50%−60% range observed in EM compost aligns with the optimal window for microbial activity. At moisture levels below 40%, microbial metabolism slows due to water-stress limitation, while levels above 65%−70% restrict oxygen diffusion and promote anaerobic conditions (Richard et al., 2002; Zhang et al., 2022). The narrower moisture range in EM treatments may reflect more uniform microbial heat production, which moderates water evaporation rates. Because supplemental water additions and evaporative losses were not quantified on a per-bin basis, this interpretation remains tentative; the narrower moisture range in the EM bins could equally reflect the higher water-holding capacity of a more humified matrix, and should be verified in future work using explicit per-bin water and evaporation balances (Richard et al., 2002).

The pH trajectory initial rise during ammonification followed by stabilization near neutrality is typical of well-managed composting (Anayet et al., 2024). The slightly higher mature-compost pH in EM bins (7.5 vs. 7.0) may reflect greater ammonia release during more intense nitrogen mineralization, though both values fall within the acceptable range for agricultural application. Read together, these physicochemical results from the present study—the earlier thermophilic onset, the narrower stabilized moisture range, and the near-neutral mature pH in the EM bins relative to the controls—indicate a consistent pattern in which inoculation primarily accelerated and tightened the maturation trajectory rather than altering the chemical endpoints, an interpretation grounded in the measured process indicators of this experiment rather than inferred solely from the cited literature.

Chemical transformations and nutrient dynamics

The faster decline in organic matter and C/N ratio in EM compost indicates that inoculation accelerated the breakdown of labile carbon substrates and promoted earlier entry into the humification phase. Nutrient enrichment during composting is driven by concentration effects as carbon is lost as CO 2 and by microbial assimilation that retains nitrogen in biomass (Kominoski et al., 2017; Brown et al., 2022). The higher total nitrogen in EM compost relative to controls (2.1% vs. 1.7%, a 24% difference) reflects greater nitrogen retention. Enhanced microbial assimilation and reduced ammonia volatilization are plausible contributing mechanisms reported elsewhere, given that EM-associated lactic acid bacteria can lower local pH and suppress NH 3 loss (Jusoh et al., 2013); however, neither microbial biomass nor nitrogen losses were measured in the present study, so these explanations remain inferential and cannot be confirmed from our data. Moreover, the higher nutrient concentrations in EM compost do not by themselves demonstrate greater absolute nutrient retention, because concentrations can increase simply through the loss of dry matter (mainly carbon as CO 2) as composting proceeds. Because per-bin dry-matter mass balances were not computed, the reported nitrogen, phosphorus, and potassium differences should be read primarily as concentration effects, and true mass-based retention remains to be confirmed; the term “retention” used here and in the abstract should be understood in this qualified sense.

Phosphorus and potassium gains of 15%−18% over controls likely reflect greater mineralization of organically bound nutrients by the more active microbial community in EM treatments. The absence of heavy-metal differences between treatments confirms that EM inoculation does not mobilize or introduce contaminants a prerequisite for safe agricultural use (Hemidat et al., 2018).

Microbial dynamics and EM effectiveness

The higher thermophilic bacterial counts and sustained microbial diversity in EM compost are consistent with EM reported ability to augment indigenous communities rather than simply replacing them (Mironov et al., 2024). The dominance of Firmicutes (Bacillus spp.) during the thermophilic phase and the shift toward Actinobacteria during cooling follow the classical microbial succession pattern in composting (Chandna et al., 2013; Ren et al., 2016), but EM inoculation shifted the relative abundances toward higher representation of functional decomposers.

qPCR data confirmed that EM-associated taxa persisted and proliferated throughout the 60-day process, with 10%−25% higher relative abundances than in controls. Such taxa are known to produce extracellular enzymes (cellulases, proteases, and lipases) that facilitate the breakdown of recalcitrant substrates such as cellulose and lignin (Huang et al., 2018; Kaiser et al., 2010), and a positive feedback between enzyme production, substrate availability, and microbial growth could in principle accelerate decomposition (Cleveland et al., 2006; Campbell et al., 2022). However, enzyme activities and substrate-degradation rates were not measured in this study; the link between the inoculated consortium and the faster decomposition kinetics is therefore inferred from the sequencing and process data rather than demonstrated directly, and should be confirmed in future work that quantifies enzyme activity.

Compost quality and potential applications

A germination index of 85% in EM compost exceeds the 80% threshold widely used to classify compost as mature and non-phytotoxic (Bazrafshan et al., 2016; Helfrich et al., 1998). The 35% increase in humic substances further confirms advanced humification, which improves the soil-amendment value of the compost by enhancing cation-exchange capacity and water-holding properties (Filcheva and Tsadilas, 2002). By contrast, control compost at 70% GI would still require additional curing before safe field application.

The nutrient profile of mature EM compost (N 2.1%, P 0.45%, K 1.2%) compares favorably with established quality standards for soil improvers (Silva et al., 2007; Al-Sari et al., 2018). Taken together, these results suggest that EM-inoculated household composting produces a product suitable for direct agricultural use, supporting smallholder farming and urban greening programs in Zanzibar.

Study limitations

Several limitations should be considered when interpreting these findings. First, the experiment was conducted at small, household scale with only three replicates per treatment, which limits statistical power and the generalizability of the results. Second, composting was carried out in bins of three different volumes (6, 9, and 21 L); because each of the three volumes was represented by only a single bin per treatment rather than being independently replicated within treatment, the effect of bin volume may be partly confounded with treatment, and the design approaches pseudoreplication rather than fully independent true replication. Third, several mechanisms invoked in the discussion were not measured directly: nitrogen losses (ammonia volatilization), microbial biomass, and extracellular enzyme activities were inferred rather than quantified, so the corresponding interpretations remain tentative. Fourth, the microbial community results are presented in summarized form; fuller relative-abundance tables, diversity plots, and qPCR standard-curve details are required to substantiate the sequencing-based conclusions, and independent validation on larger sample sets would strengthen them. Finally, the study covered a single composting cycle at one site and in one season; longer-term, multi-site, and multi-season trials are needed before the results can be generalized. These limitations do not negate the observed benefits of EM inoculation, but they define the boundaries within which the present conclusions should be read and identify clear priorities for future work. In addition, the design did not include a molasses-water carrier control; because molasses is itself a labile carbon source that can stimulate microbial activity, part of the response attributed to EM inoculation may derive from the carrier rather than the microbial consortium, and a molasses-only control should be incorporated in future trials. Because the largest vessel was only 21 L, the results should be treated as preliminary and require validation in community- and industrial-scale systems before wider application. The available dataset contains pooled treatment summaries rather than recoverable raw results for each 6, 9, and 21 L vessel; cross-scale consistency therefore cannot be assessed and no per-volume values have been reconstructed retrospectively (Supplementary Table S1). Finally, exact container dimensions and hole counts, the EM product and batch identity, the activation temperature and viable concentration, and quantitative turning and watering logs were not retained. These omissions limit reproducibility and are reported transparently rather than filled with unsupported estimates.

EM inoculation of household organic waste significantly improved composting performance across physical, chemical, and biological indicators. Compared with uninoculated controls, EM treatments reached higher thermophilic temperatures sooner, maintained more stable moisture, achieved faster organic matter decomposition and nutrient concentration, produced higher microbial diversity, and yielded a more mature and less phytotoxic final product. Heavy metals remained within safe limits regardless of treatment. The principal scientific advance of this work is the demonstration, within a single controlled experiment on mixed household waste, that EM inoculation accelerates the maturation trajectory (earlier and longer thermophilic heating, faster organic-matter and C/N decline, and higher compost maturity) while simultaneously enriching EM-associated taxa quantified by qPCR, an integrated physicochemical-molecular characterization not previously reported for domestic organic waste in this setting. These benefits should nonetheless be interpreted in light of the design limitations detailed in Section 4.5.

These results support the use of EM-inoculated composting as a low-cost, decentralized strategy for managing household organic waste in Zanzibar and similar settings. Scaling this approach could reduce landfill pressure, lower greenhouse gas emissions from open dumping, and provide nutrient-rich soil amendments for local agriculture. Future work should evaluate the long-term effects of EM-compost application on soil health, crop yield, and emissions under field conditions, and should assess the economic feasibility and community acceptance of household-scale EM composting programs.