Greenhouse Gas Emissions and Nutrient Recovery from Fish Waste During Composting and Burial

Abstract

Fish-processing residues represent a significant environmental challenge due to their high moisture and nitrogen contents, which favor greenhouse gas (GHG) emissions during degradation. This study evaluated how different waste management strategies affect GHG emissions from fish waste, including conventional composting (Bulk), composting amended with biochar (BulkBioch), burial with soil (S), and burial with soil plus sawdust (BulkS). Daily emissions of CH₄, N₂O, and CO₂ were monitored, and cumulative emissions were modeled using generalized additive models. Composting treatments (Bulk and BulkBioch) released higher CO₂, suggesting greater microbial degradation, while burial treatments developed earlier anaerobic conditions with reduced decomposition efficiency. Bulk showed the highest cumulative CH₄ and CO₂ emissions, whereas N₂O fluxes were greater in burial methods, reaching 2.18 g N₂O kg⁻¹ TS in S. Biochar addition was associated with 15% and 10% lower CH₄ and N₂O emissions, respectively, and earlier stabilization of CH4 emissions. In global warming potential, BulkBioch presented the lowest climate impact (305 g CO₂-eq kg⁻¹ fish), followed by Bulk (338 g CO₂-eq kg⁻¹), whereas BulkS reached up to 599 g CO₂-eq kg⁻¹. The use of bulking agents in burial resulted in lower CH₄ buildup and greater nutrient retention. Overall, combining bulking agents and biochar may represent a promising strategy to mitigate GHG emissions while supporting nutrient conservation.

1. Introduction

The fish industry is projected to reach 194 million tons of production by 2026 [1], which will generate vast amounts of waste, particularly during filleting, a stage that accounts for an average of 70% of the carcass [2]. Fishmeal production is one of the most profitable and common valorization strategies for these by-products; however, constraints such as the need to discard lots condemned for sanitary reasons and the limited storage capacity within processing facilities often undermine this option [3]. Therefore, alternative technologies are required to ensure proper disposal, aiming to minimize pathogen dissemination and mitigate air, soil, and water contamination.

Composting is an efficient, low-cost, and simple management strategy that eliminates pathogens due to the high temperatures reached [4,5], in addition to generating a nutrient-rich organic fertilizer [6]. In contrast, burial is still used as a disposal option in aquaculture units due to its operational simplicity and low cost, but it poses environmental risks such as pathogen dissemination and potential groundwater contamination [7]. Both disposal methods contribute to greenhouse gas (GHG) emissions, resulting in significant carbon losses as CO₂ and CH₄ [8], as well as nitrogen losses as N₂O [9].

The use of additives has shown potential to mitigate GHG emissions and improve degradation conditions. Bulking agents such as sawdust enhance aeration, adjust the C:N ratio, and supply lignin for humic compound formation [10]. Biochar, in turn, is recognized for its ability to adsorb gases such as CO₂ and NH₃ and to reduce the formation of anaerobic pockets [11]. Experimental evidence supports these benefits: [12] reported that granular biochar increased porosity by 4.02% and decreased CH₄ emissions by 22.15% in pig manure composting with wheat straw, while [13] found that the addition of 10% rice straw biochar reduced cumulative N₂O emissions by 55% during the initial stage of agricultural waste composting.

Excessive gaseous losses may also impair the agronomic quality of the final compost. High nitrogen losses reduce the availability of this key macronutrient [14], while reduced phosphatase activity can diminish phosphorus concentrations [15]. Furthermore, recent studies indicate that potassium and other nutrients may also be affected under poorly managed composting conditions, reducing the fertilizing potential of the final product [16,17].

Evaluating how different fish waste disposal strategies influence GHG emissions during organic matter degradation is essential for improving the environmental sustainability of aquaculture waste management. Despite growing interest in alternative treatments such as composting with additives, limited information is available regarding the temporal patterns of GHG emissions from fish waste under different management approaches. Therefore, the present study aimed to investigate the dynamics of GHG emissions during the degradation of fish waste managed through composting systems (with and without biochar) and burial systems (with and without bulking agents). In addition, the study examined how these management strategies influence key process parameters, including temperature and pH throughout the decomposition process, as well as the nutrient composition of the final compost.

This study provides a novel and integrated evaluation of GHG emissions and nutrient recovery from fish waste under the main disposal strategies currently adopted in aquaculture systems, particularly in high fish-producing countries such as Brazil. These management practices were selected to represent realistic scenarios observed in productive settings. Thus, beyond comparing treatments, this study sought to understand how each management strategy influences gas emissions, process dynamics, and nutrient recovery under practical conditions. This approach advances the literature and supports circular economy principles by promoting the sustainable valorization of fish-processing residues.

2. Materials and Methods

2.1. Experimental Site and Waste Characterization

The experiment was conducted in a greenhouse at the experimental area of Embrapa Western Agriculture, located in Dourados, Brazil (22°16′30″ S, 54°49′00″ W, at an altitude of 408 m). The region’s climate, as classified by the Köppen system, is Cwa—a Humid mesothermal climate, characterized by hot summers and dry winters, with an annual precipitation of 1500 mm and an average annual air temperature of 22 °C.

Fish waste was collected fresh at the time of filleting from a commercial processing unit in Itaporã, MS, and consisted of heads, scales, carcasses, and whole fish. The bulking agent (sawdust) was obtained from a local sawmill located near Dourados, and soil was collected at the experimental site. Biochar was produced from bamboo biomass. The composition of the raw materials is presented in

Table 1. Chemical composition of raw materials used in the composting and burial of fish waste.

Parameters 1/ Raw Materials	Fish Waste	Sawdust	Biochar	Soil
pH	6.20	4.46	8.24	6.08
TS (%)	27.81	81.54	82.37	89.00
VS (% TS)	80.60	98.55	89.58	13.02
C (% TS)	48.35	46.47	49.93	9.22
N (% TS)	5.39	0.14	0.63	0.14
C:N	8.97	331.9	79.25	65.85
EE (% TS)	38.80	0.56	-	-
NDF (% TS)	-	88.89	71.29	-
ADF (% TS)	-	73.12	64.06	-
Cellulose (% TS)	-	30.65	21.64	-
Hemicellulose (% TS)	-	12.92	15.08	-
Lignin (% TS)	-	42.82	40.88	-
SA (m2 g−1)	-	-	7.45	-
PS (nm)	-	-	7.55	-

¹ TS, total solids; VS, volatile solids; NDF, neutral detergent fiber; ADF, acid detergent fiber; pH, hydrogen ion concentration; EE, ethereal extract; SA, surface area; PS, pore size.

. The surface area and pore size data were obtained from [18].

The experiment was designed as a pilot-scale study to represent realistic operational scenarios of fish waste management. Due to logistical, economic, and biosafety constraints inherent to large-scale composting and burial systems, each treatment was conducted in a single experimental unit. Nevertheless, the system was intensively monitored over time, and repeated measurements were used to characterize the temporal dynamics of greenhouse gas emissions and process parameters.

Fish waste was placed in four masonry cells, each with an area of approximately 1 m² and an estimated capacity of 300 kg of fresh material. The fish waste was added to the cell in alternating layers, as follows: (i) bulking agent and fish waste (Bulk); (ii) bulking agent and fish waste with 10% biochar on a fish waste mass basis (BulkBioch); (iii) bulking agent and fish waste buried (BulkS), and (iv) fish waste buried (S) (Figure 1).

For composting piles, materials were layered alternately within the cells, starting with sawdust at the base and finishing with sawdust at the top (Figure 1). The bulking agent: fish waste ratio was 1:2.5 (mass:mass), and the biochar was placed over the fish waste layer and distributed uniformly. Under the burial condition, the base soil was lined with a layer of bulking agent, followed by an additional soil layer, and approximately 150 kg of fish waste was added. For covering, one cell received a bulking agent followed by soil (BulkS), whereas the other was covered with soil only (S) (Figure 1).

In composting, forced aeration was applied during the first six weeks using an air blower at a continuous flow of 0.6 L kg⁻¹ VS min⁻¹, as recommended by [19]. Furthermore, three turnings were conducted at 50, 70, and 90 days of composting. The material was removed from the composting cell and placed on a plastic canvas for homogenization, sample collection, and moisture content adjustment. No management was applied to burial treatments to avoid process interference. During the burial period, simulated rainfall totaled 301.3 mm, based on the cumulative precipitation recorded at the experimental site throughout the trial. Water was evenly applied to all treatments according to the rainfall schedule to mimic natural precipitation conditions. The maintenance of soil moisture through rainfall simulation likely affected microbial activity, organic matter degradation, nutrient transformations, and gas fluxes.

Both the composting and burial processes took 135 days, during which temperature was monitored continuously using dataloggers. Temperature readings were taken from the base, center, and top of the cells to calculate the average temperature. The levels of total solids (TS), volatile solids (VS), and pH were determined in the samples collected from the initial material and at 135 days of composting. The TS and VS content was measured using the methodology outlined in [20], while the pH was determined using the method described in [21]. To assess the quality of the compost, the levels of macrominerals (N, P, and K) were determined using an Inductively Coupled Plasma Atomic Emission Spectrometer (ICP-OES), PerkinElmer, Optima 8300 model (Dual View). C and N concentrations were determined using a VARIO MACRO model Elemental Analyzer [22]. The ether extract contents were determined using the Randall method (INCT-CA G-005/1) as outlined in [23]. NDF, ADF, cellulose, hemicellulose, and lignin analyses were performed as previously described [24].

2.2. Quantification of CH₄, N₂O, and CO₂

Fluxes of CH₄, N₂O, and CO₂ were quantified using the closed static-chamber technique. The chamber was an opaque rectangular enclosure (73.2 L). Emissions were monitored during 135 days at variable intervals of 2–7 days, totaling 45 deployments per windrows and 37 deployments for the burial treatments. For each deployment, the chamber was seated on the cell’s sampling interface, the headspace air was mixed with an internal fan for 30 s, and the headspace temperature was recorded with a digital thermometer attached to each chamber. Gas samples were withdrawn with 20 mL syringes at 3, 6, 9, and 12 min after closure (four samples per deployment). Sampling started at 8:00 local time to approximate daily average soil N₂O and CH₄ emissions [25].

Samples were analyzed by gas chromatography using a Thermo Scientific Trace 1310 gas chromatograph equipped with Porapak Q/N columns, dual detectors (electron-capture detector, ECD; and flame-ionization detector, FID), and a TriPlus RSH autosampler for automatic injection. Before each run, certified standards with known concentrations were injected to construct calibration curves used to convert peak areas to gas concentrations in samples.

Gas concentrations were temperature-corrected to the headspace temperature measured at the time of sampling. Chamber fluxes were computed from the linear rate of change in headspace concentration over time following [26]:

F = δC/δt × 60 × (V/A) × M/Vm,

(1)

where F represents the gas flux (mg m⁻² h⁻¹); δC/δt is the temporal change in gas concentration within the chamber during the closure interval (ppm min⁻¹), V is chamber volume (m³), A is the covered surface area, M is the molecular weight of the CH₄ (16 g mol⁻¹), CO₂ (44 g mol⁻¹), N₂O (44 g mol⁻¹); Vm is the molar volume corrected to sampling conditions:

Vm = 0.02241 × ((273,15 + T)/273.15) × p0/p1,

(2)

with Vm in m³ mol⁻¹ (0.02241 m³ mol⁻¹ = 22.41 L mol⁻¹), T the chamber headspace temperature (°C), p0 standard atmospheric pressure at sea level, and p1 the site atmospheric pressure, estimated from the barometric equation, site altitude.

Cumulative emissions were estimated using the trapezoidal integration method, which assumes a linear variation between two consecutive sampling points. For each gas (CH₄, N₂O, and CO₂), the daily flux values (expressed per unit area or per unit mass of dry waste) were interpolated using the average of two successive measurements multiplied by the time interval between them [27].

2.3. Global Warming Potential (GWP)

The global warming potential (GWP) of each treatment was estimated from the cumulative emissions of methane (CH₄) and nitrous oxide (N₂O). Only CH₄ and N₂O were considered in this calculation because carbon dioxide (CO₂) emissions were of biogenic origin and therefore were not included in the GWP assessment. Emissions were converted to carbon dioxide equivalents (CO₂-eq) using global warming potential factors of 27 for CH₄ and 273 for N₂O over a 100-year time horizon, according to the Intergovernmental Panel on Climate Change [28].

Total GWP was calculated as the sum of the CO₂-equivalent contributions of both gases and expressed per kilogram of initial fish dry matter to enable comparison among treatments. The calculation was performed as follows:

GWP = (CH₄ × 27) + (N₂O × 273),

(3)

where CH₄ and N₂O represent the cumulative emissions of methane and nitrous oxide, respectively, expressed per kilogram of initial fish dry matter.

2.4. Thermal (TGA–DSC) Analysis

Samples for thermal analysis were collected on day 135 during the composting and burial process. Thermal analyses (TGA-DSC) were performed on a Netzsch STA449 F3 Jupiter^® thermogravimetric analyzer using approximately 5.0 mg subsamples and α-alumina crucibles. The purge gas flow rate (air) was set to 50 mL min⁻¹, and the heating rate was conducted at 20 °C min⁻¹ over a temperature range of 30 to 1000 °C [29].

2.5. Plant Growth Assay

A germination test was conducted using lettuce seedlings grown with the organic fertilizers produced in the experiment. Among the products resulting from the evaluated fish waste disposal methods, only those derived from composting were selected for this assessment. This selection was based on their exposure to an intense, sustained thermophilic phase, which led to greater reductions in solid constituents and ensured adequate stabilization and sanitization of the material. Additionally, these composted fertilizers exhibited higher concentrations of plant-essential nutrients, particularly nitrogen (N), phosphorus (P), and potassium (K), compared to the materials obtained from the burial treatments.

The bio-compound assay was performed in a completely randomized design with a 2 × 4 factorial arrangement, comprising two organic fertilizers (with and without biochar) and four application rates (0, 25, 50, and 100 g), with four replicates per treatment. Each experimental unit consisted of a 3 L pot containing 2 kg of unfertilized soil and one lettuce seedling (approximately 8 g fresh mass and 9 cm in height). Application rates were defined based on typical field recommendations for organic fertilizers in vegetable production systems (up to 60 t ha⁻¹; [30]) and proportionally adjusted to the pot scale. The soil was collected from an area with no history of fertilization. Pots were irrigated daily for 30 days, after which plants were harvested to determine shoot fresh weight, shoot length, root fresh weight, and root length.

2.6. Statistical Analysis

The experimental design consisted of one composting cell per treatment; therefore, no spatial replication was performed. Consequently, the variability presented in the emission curves reflects temporal variation in gas fluxes throughout the monitoring period rather than experimental replication. Time was treated as a repeated-measures dimension within each treatment, and statistical modeling focused on characterizing temporal emission dynamics independently for each management strategy. Accordingly, analyses of the composting experiment were descriptive and exploratory, whereas inferential statistics were reserved for the replicated germination assay.

Statistical modeling of daily and cumulative gas emission trends for each gas and fish waste treatment was done following three steps: (i) calculation of daily gas emissions on a per dry manure basis (g d⁻¹ kg⁻¹ dry fish waste) from emission 12 min dataset; (ii) estimation of mean daily fluxes using generalized additive models [31] fitted to the daily time series data, with thin plate regression splines as the default smoothing function in the “gam” function of the “mgcv” package in R version 4.5.3 [32]; and (iii) to propagate uncertainties in daily emissions, Monte Carlo methods were used to sample 10,000 vectors of parameter values from the variance–covariance matrix of parameter estimates [32], simulating 10,000 daily emission curves consistent with our data. From these curves, we computed the mean and 95% confidence intervals of daily gas emissions. This approach follows the methodology applied by [33], allowing for a flexible representation of the non-linear variation in emissions over time within each treatment.

An exploratory correlation analysis was conducted to investigate potential associations between the final nutrient concentrations in the compost (N, P, and K) and the thermal behavior of the piles during composting. For each treatment, the final values of N, P, and K were extracted from the last sampling point. Correlation coefficients were estimated using both Pearson’s (linear correlation) and Spearman’s (rank-based correlation) methods provided in Appendix A (

Table A1. Exploratory Pearson’s correlation coefficients between final nutrient concentrations (N, P, K; n = 4, one value per treatment) and thermal descriptors of the composting process (T_mean, T_max, T_sd, T_AUC; calculated from daily records).

	N_final	P_final	K_final	T_final	T_max	T_sd	T_AUC
N_final	1.000	0.962	0.996	0.902	0.919	0.968	0.903
P_final	0.962	1.000	0.979	0.973	0.975	0.945	0.973
K_final	0.996	0.979	1.000	0.919	0.931	0.956	0.920
T_mean	0.902	0.973	0.919	1.000	0.999	0.940	1.000
T_max	0.919	0.975	0.931	0.999	1.000	0.957	0.999
T_sd	0.968	0.945	0.956	0.940	0.957	1.000	0.941
T_AUC	0.903	0.973	0.920	1.000	0.999	0.941	1.000

and

Table A2. Exploratory Spearman’s correlation coefficients between final nutrient concentrations (N, P, K; n = 4, one value per treatment) and thermal descriptors of the composting process (T_mean, T_max, T_sd, T_AUC; calculated from daily records).

	N_final	P_final	K_final	T_final	T_max	T_sd	T_AUC
N_final	1.0	0.8	1.0	0.8	0.8	0.6	0.8
P_final	0.8	1.0	0.8	1.0	1.0	0.8	1.0
K_final	1.0	0.8	1.0	0.8	0.8	0.6	0.8
T_mean	0.8	1.0	0.8	1.0	1.0	0.8	1.0
T_max	0.8	1.0	0.8	1.0	1.0	0.8	1.0
T_sd	0.6	0.8	0.6	0.8	0.8	1.0	0.8
T_AUC	1.0	0.8	1.0	0.8	0.8	0.6	0.8

). Given the experimental design, which comprised four treatments without replication, the analysis was restricted to descriptive and exploratory purposes. Accordingly, no statistical inference was applied, and the results are presented as trends that may indicate possible relationships between nutrient retention, organic matter degradation, and process temperature dynamics.

To further integrate daily gas emissions and temperature dynamics, a Principal Component Analysis (PCA) was performed on the daily dataset (CH₄, N₂O, CO₂, and temperature), enabling the visualization of multivariate trajectories for each treatment. Temperature evolution during composting is shown in Figure A1 (Appendix B). Treatments were projected as temporal trajectories, and k-means clustering was applied to daily points, identifying three clusters corresponding to early, thermophilic, and late composting phases. This multivariate analysis was exploratory in nature and intended to describe temporal progression and emission–temperature coupling rather than to test statistical differences among treatments. The resulting multivariate patterns are illustrated in Figure A1, while detailed loading values and correlation matrices used for the PCA are provided in Appendix A (

Table A3. Summary of the PCA approach applied to daily gas emissions and temperature dynamics.

Step	Description
Data input	Daily dataset including CH4, N2O, CO2 fluxes and mean pile temperature (T_treat).
Scaling/centering	Variables standardized (z-score) before PCA.
PCA method	Performed using prcomp function in R with correlation matrix
Retained components	The first two PCs were retained for visualization.
Trajectory analysis	PCA scores plotted as temporal trajectories per treatment.
Cluster identification	K-means clustering applied to PCA scores (k = 3).
Phase interpretation (exploratory)	Clusters interpreted as early, thermophilic, and late composting phases (descriptive only).
Software	R 4.5.3 (R Core Team), packages: stats, ggplot2, factoextra.

).

Germination test data were subjected to analysis of variance (ANOVA), and when the F-test indicated significant treatment effects, means were compared using Tukey’s test at the 5% significance level (p < 0.05).

A generative artificial intelligence tool (ChatGPT-5, OpenAI, San Francisco, CA, USA) was used as technical support for troubleshooting and refining R scripts related to data processing and graphical visualization. Its use was limited to identifying coding issues, suggesting syntactic corrections, and improving the formatting of figures and plot elements. No AI tools were used for statistical decision-making, hypothesis testing, or interpretation of results.

2.7. Limitations of the Study

The composting and burial experiment was conducted under pilot-scale conditions with a lack of replication, since each waste-management treatment consisted of a single experimental unit. Therefore, the observed differences in gas emissions, process dynamics, and compost characteristics should be interpreted as exploratory trends rather than statistically confirmed treatment effects, with limited generalizability to commercial-scale systems. Further replicated studies under full-scale operational conditions are needed to validate the trends observed here.

3. Results and Discussion

3.1. Temperature and pH Dynamics

Composting cell temperatures (Figure 2 and Figure A1) remained within the thermophilic range (above 45 °C) under both conditions (with and without biochar) for most of the process. The cell without biochar addition stayed in this phase for 87 days, with an average thermophilic temperature of 53 °C. In contrast, the cell amended with 10% biochar remained above 45 °C for 105 days, with a mean temperature of 54.1 °C. The peak temperature of both cells was reached at day 100, corresponding to 64.3 °C in the control and 69.6 °C in the treatment with 10% biochar. In the buried waste, temperatures remained below 45 °C throughout the process in both cover conditions. The maximum temperatures in the buried residues occurred at days 102 and 121, reaching 40.8 °C in the residues covered only with soil and 48.0 °C in those covered with sawdust plus soil, respectively.

The persistence of thermophilic temperatures during composting was likely associated with management practices such as forced aeration, pile turning, and moisture adjustment. These practices are essential to establish favorable conditions for microbial activity, ensuring sufficient oxygen supply, reduced compaction, and availability of organic substrates [4]. In addition to management strategies, environmental conditions such as high ambient temperature and relative humidity also contributed to maintaining elevated composting temperatures. The inclusion of 10% biochar likely enhanced metabolic heat generation within the composting mass, as biochar can store heat within its porous structure, thereby intensifying organic matter decomposition [8].

Temperature peaks reflected the oxidation of organic matter by aerobic microorganisms. During this process, heat is released as a natural by-product of exothermic reactions, while organic compounds are converted into simpler end-products such as CO₂ and H₂O [34,35].

In contrast, the burial method, where residues were arranged in a single layer, likely caused significant compaction that slowed degradation. Soil coverage further reinforced this compaction, particularly during water addition events [36]. Moreover, the absence of management interventions under this method contributed to the observed temperature profile. Similar results were reported by [37] when whole swine carcasses were buried in soil. The use of sawdust as an additional cover material for fish residues increased the temperature, likely by enhancing oxygen penetration and circulation within the buried mass, thereby favoring microbial development.

All pH values recorded throughout the experiment remained within the range considered optimal for microbial degradation processes. The pH exhibited a similar overall trend in both composting and burial treatments, with an initial decrease followed by a gradual increase toward neutrality or slight alkalinity. In composting, initial values ranged from 5.33 to 6.30 and increased to 7.23 and 8.35 in the treatments without and with biochar, respectively. In burial, values varied within a narrower range, from 6.15–7.0 initially to 6.35–6.82 at the end of the process, remaining close to neutrality. This trend reflects the typical dynamics of organic matter decomposition. The initial acidification is mainly associated with the accumulation of CO₂ and volatile fatty acids during the rapid decomposition of easily degradable substrates [13]. As degradation progresses, these acids are consumed, and ammonium is formed, leading to a gradual increase in pH.

Although biochar addition resulted in a slightly higher final pH compared to the control (8.35 vs. 7.23), the magnitude of this increase was less pronounced than reported in previous studies. For instance, Ref. [38] observed a final pH of 9.12 when 20% wood biochar was added to food waste composting. Ref. [39] emphasized the alkalinizing effect of biochar due to its high content of basic cations (K⁺, Na⁺, Mg²⁺, Ca²⁺) and carbonate species. The smaller increase observed here may be attributed to the specific experimental conditions, including the type of waste, biochar proportion, and moisture management, which likely moderated the buffering capacity of biochar.

3.2. Daily CH₄, N₂O, and CO₂ Emissions

In the static windrows, the highest CH₄ emission was recorded on day 24 in the treatment with biochar (BulkBioch; 98.38 mg m⁻² h⁻¹) and on day 58 in the treatment without biochar (Bulk; 129.98 mg m⁻² h⁻¹) (Figure 3). In the burial treatments, the greatest CH₄ emissions occurred during the first weeks, with a peak of 32.5 mg CH₄ m⁻² h⁻¹ when residues were covered exclusively with soil (S), and values below 27 mg CH₄ m⁻² h⁻¹ in the treatment covered with sawdust (BulkS), stabilizing at around 5 mg CH₄ m⁻² h⁻¹ after 72 days (Figure 3).

Overall, CH₄ emissions in all treatments decreased over the course of the study, a typical pattern associated with the intensity of aerobic decomposition (Figure 3). The period of more intense methane release usually coincides with the anaerobic phase of the process. In static windrows, residues tend to compact, hindering oxygen penetration, favoring the development of methanogenic microbial communities, and consequently leading to higher CH₄ emissions [33]. Some factors, such as the inclusion of biochar, aeration, and pile turning, help maintain aerobic conditions and mitigate GHG emissions, including CH₄ [17,40,41]. The addition of biochar to composting can increase oxygen concentration and control methanogenic activity, reducing GHG emissions by 27–32% [42,43,44].

In this study, biochar increased CH₄ emissions during the first 50 days of composting (Figure 3). This effect likely resulted from the layered application of the additive over fish residues, which likely promoted anaerobic microsites, especially after moisture adjustment. Water may have occupied biochar pores, reducing its aeration capacity and limiting oxygen transfer. Similar results were reported by [45], who observed a higher initial CH₄ emissions with 5% biochar, attributed to greater organic matter availability. Likewise, the high organic matter content of fish residues in this study (Table 1) may have contributed to the elevated CH₄ emissions in the initial phase. This interpretation is supported by [46], who reported that high availability of readily degradable organic matter, combined with rapid microbial oxygen consumption, promotes the formation of anaerobic zones and increases CH₄ emissions. After pile turning, biochar was homogenized with the other components, eliminating the layered arrangement and allowing its beneficial effects to prevail.

Covering residues exclusively with soil resulted in lower CH₄ emissions (Figure 3). This lower emission may be related to the lower moisture observed in this treatment, which likely limited the formation of anaerobic microsites and, consequently, constrained the conditions required for methanogenesis. In addition, restricted gas diffusion through the soil layer, together with CH₄ oxidation in partially oxygenated microsites, may have further decreased net emissions [47].

Regarding N₂O, composting (Bulk and BulkBioch) showed more irregular emissions and sharper peaks compared to the burial treatments (Figure 4). In the static windrows, the initial peaks occurred at day 14, with values of 55.80 mg m⁻² h⁻¹ for Bulk and 15.12 mg m⁻² h⁻¹ for BulkBioch, while during the same period emissions were 9.02 mg m⁻² h⁻¹ for S and 10.13 mg m⁻² h⁻¹ for BulkS (Figure 5). All residue management strategies exhibited a similar pattern of delayed N₂O emission. The highest N₂O peaks occurred near the end of the process, reaching 82.32 and 68.06 mg m⁻² h⁻¹ at days 125 and 132 for the sawdust-plus-soil and soil-only covers, respectively, and corresponding to 135.68 and 165.65 mg m⁻² h⁻¹ at day 125 in the Bulk and BulkBioch treatments, respectively (Figure 4).

The delayed N₂O emission observed in this study agrees with [48], who reported higher emissions at the final stage of rice straw composting. This pattern likely reflects the combined influence of temperature and oxygen limitation on N₂O emissions. In the present study, although thermophilic conditions persisted, the temperature decline after day 100 (Figure 2) coincided with increased N₂O emissions, suggesting the reactivation of microbial processes associated with nitrification and denitrification. High temperatures can inhibit nitrification, limiting N₂O production during the thermophilic phase [49,50]. In contrast, other studies report early N₂O peaks. Peng et al. [5] observed initial emissions linked to high microbial activity, readily degradable substrates, moderate temperatures, and heterogeneous oxygen distribution. These conditions favor nitrification and denitrification, increasing N₂O emissions. Similarly, early N₂O release has been attributed to NH₄⁺ oxidation to NO₂⁻ [51,52,53].

The daily CO₂ emissions exhibited contrasting behaviors between composting (Bulk and BulkBioch) and burial (BulkS and S) of fish residues. In the composting treatments, multiple peaks were observed throughout the process, with the most pronounced increases occurring after pile turnings (Figure 5). Both treatments reached their maximum at day 52, with values of 338,911.0 and 256,337.7 mg m⁻² h⁻¹ for BulkBioch and Bulk, respectively (Figure 5). In the burial treatments, the highest emissions occurred on day 2 (175,613.7 mg m⁻² h⁻¹ in S and 123,844.0 mg m⁻² h⁻¹ in BulkS), followed by a sharp decline and lower-amplitude fluctuations (Figure 5).

CO₂ emission is widely recognized as an indicator of the intensity of aerobic microbial respiration during the decomposition of organic matter [17]. More pronounced peaks reflect higher degradation rates, associated with an efficient biological process [45]. In the present study, the highest CO₂ fluxes observed in the composting treatments coincided with the periods of turning and with forced aeration during the first 50 days of the process. As discussed previously, turning and aeration likely sustained aerobic conditions, contributing to higher CO₂ fluxes.

The presence of biochar and sawdust in the composting of fish waste may have enhanced this effect. As noted earlier, sawdust may improve physical structure and aeration of the matrix while also supplying additional carbon, which may contribute to CO₂ emissions. The results are consistent with [11], who reported a 7.4% increase in CO₂ emission when poultry manure compost received the highest dose of bamboo biochar (10%). Other studies have reported that higher CO₂ emissions in composting with biochar addition may result from the abiotic oxidation of biochar or from the labile carbon fraction of biochar, which serves as an energy source for microorganisms [35,54].

In the burial treatments, the absence of turning and forced aeration rapidly limited oxygen diffusion, promoting the early establishment of anaerobic conditions. This limitation reduced the activity of aerobic microorganisms, resulting in lower emissions and a slower decomposition process. The more intense CO₂ release observed at the beginning of the burial treatments, under both conditions, may represent a response to the consumption of residual oxygen. Subsequently, due to compaction, the environment shifted toward anaerobiosis, thereby reducing emissions.

To complement the univariate assessment of daily gas fluxes, a multivariate analysis was performed to integrate emission dynamics with temperature over time. In addition to the temporal dynamics of individual gases, multivariate analysis of daily emissions and temperature revealed clear treatment-dependent trajectories (Figure 6 and Figure A1, Appendix B). The PCA separated the composting treatments (Bulk and BulkBioch) from the burial treatments (S and BulkS), with clusters corresponding to distinct composting phases. Bulk and BulkBioch exhibited longer trajectories, reflecting the transition across multiple stages, including a sustained thermophilic phase characterized by high CO₂ fluxes and active microbial degradation. In contrast, the burial treatments showed shorter trajectories and clustered in early-phase regions, consistent with lower temperatures and more limited aerobic activity.

3.3. Generalized Additive Model and Cumulative Emissions

Considering the results fitted by the generalized additive model (GAM), daily GHG emissions, expressed per kilogram of TS fish waste, showed distinct temporal patterns across treatments, reflecting the influence of waste management strategy, as well as the use of biochar and bulking agent, on the dynamics of organic matter degradation (Figure 7). In general, CH₄ emissions were more pronounced during the early to intermediate stages of the process, with peaks observed between 30 and 50 days of composting or burial (Figure 7). N₂O emissions, in turn, varied depending on the treatment, with peaks occurring later in the process, whereas CO₂ exhibited a more constant behavior over time, though with differing intensities among treatments.

The cumulative emissions of GHG per kilogram of TS fish waste varied substantially among treatments (Figure 7). The Bulk treatment tended to show higher accumulated values of methane (0.72 g CH₄ kg⁻¹ dry waste) and carbon dioxide (2.57 kg CO₂ kg⁻¹ dry waste), whereas the highest accumulations of nitrous oxide were observed in the burial treatments S (2.18 g N₂O kg⁻¹ dry waste) and BulkS (2.08 g N₂O kg⁻¹ dry waste). The treatment with biochar (BulkBioch) presented comparatively lower cumulative emissions of CH₄ (0.61 g kg⁻¹ dry waste) and N₂O (1.06 g kg⁻¹ dry waste) compared with the Bulk treatment, representing mitigation of 15% and 10%, respectively. The incorporation of sawdust in the burial treatment positively influenced the N₂O emission profile, resulting in a value 4% lower than that recorded in the soil-only treatment (S).

Cumulative data indicate that biochar promoted earlier stabilization of CH₄ emissions (Figure 7). By day 73, 70% and 91% of total methane had been released in the Bulk and BulkBioch treatments, respectively. The faster accumulation in the BulkBioch may be associated with its higher moisture content (30.5% vs. 24.9%), which likely favored anaerobic microsites and accelerated methanogenesis in the initial phase. Later, this moisture pattern was reversed, with higher moisture content in Bulk (38.8%) than in Bulkbioch (32.5%). If confirmed as an indicator of compost maturity, this stabilization could offer operational advantages, such as faster pile turnover and reduced land use for waste treatment. It may also provide environmental benefits by reducing reliance on synthetic fertilizers.

The mitigation of N₂O emissions observed with biochar addition in the present study is consistent with [55,56,57,58], who reported a reduction of approximately 17.5% to up to 70% in cumulative emissions during composting. In contrast, Ref. [59] reported a significant increase in cumulative N₂O emissions following the addition of 10% tobacco-derived biochar during food waste composting. This discrepancy likely reflects differences in substrate characteristics, biochar properties, and process conditions [17,57,58]. The biochar used in the present study exhibited a surface area of 7.45 m² g⁻¹ and a pore size of 7.55 nm, indicating a mesoporous structure that may facilitate gas diffusion and contribute to aeration within the compost matrix. These characteristics align with reports that biochar porosity and surface area influence water retention, aeration, and microbial activity, affecting nitrogen transformation [60]. Furthermore, previous studies have shown that biochar can mitigate N₂O emissions by altering nitrogen transformation pathways and microbial activity [58].

In the present study, cumulative N₂O emissions per unit of TS residue were higher than those of CH₄, which can be explained by the high nitrogen concentrations in fish waste and the nitrogen transformation pathways discussed above. According to [17], this prominent contribution of N₂O is linked to the high availability of nitrogen, particularly in the ammoniacal form (NH₄⁺), which serves as a substrate for nitrification and denitrification. Moreover, the physical environment of composting under intermittent aeration fosters the development of microaerobic zones, which are conducive to partial nitrogen conversion into N₂O.

The more intense cumulative N₂O emissions observed in the burial treatments (S and BulkS) can be explained by a combination of physical, chemical, and microbiological factors that favor incomplete nitrification and partial denitrification, the main pathways of N₂O production [42]. This finding is particularly relevant since nitrous oxide has a global warming potential per molecule approximately 296 times greater than that of carbon dioxide (CO₂), making it one of the most impactful greenhouse gases [48]. In addition, N₂O can be transported to the stratosphere, where it contributes to ozone layer depletion, posing substantial risks to ecosystems and human health [61]. Thus, the choice of waste treatment method becomes a strategic factor in mitigating these emissions, especially given the evidence that burial favors the release of this pollutant.

Fish residues are characterized by high moisture and nitrogen contents, conditions that may favor CH₄ and N₂O formation during degradation. Within the conditions of this pilot-scale study, biochar amendment was associated with lower cumulative emissions of both gases. Further studies under replicated and commercial-scale conditions are needed to confirm these trends and to evaluate alternative methods for homogeneous biochar incorporation.

3.4. Global Warming Potential

Table 2. Cumulative CH₄ and N₂O emissions, emission intensity, and global warming potential (GWP) per kilogram of initial fish dry matter under different fish waste management treatments.

Treatment	Cumulative CH4 (g)	Cumulative N2O (g)	Emission Factor CH4 (g kg−1 Fish DM)	Emission Factor N2O (g kg−1 Fish DM)	GWP (kg CO2-eq kg−1 Fish DM)
Bulk	52.3	85.1	0.72	1.17	339
BulkBioch	47.9	83.3	0.61	1.06	305
BulkS	7.9	88.0	0.19	2.18	599
S	8.2	81.7	0.21	2.08	575

summarizes cumulative emissions and global warming potential (GWP) values across treatments, allowing direct comparison of their environmental performance. The GWP differed among treatments (Table 2), based on cumulative CH₄ and N₂O emissions expressed as CO₂-equivalent. Composting amended with biochar (BulkBioch) showed the lowest GWP (305 g CO₂-eq kg⁻¹ fish), followed by conventional composting (Bulk) (338 g CO₂-eq kg⁻¹ fish). In contrast, burial treatments were associated with higher climate impacts, reaching 599 and 575 g CO₂-eq kg⁻¹ fish for the sawdust + soil (BulkS) and soil-only (S) treatments, respectively.

The addition of biochar resulted in a 33 g CO₂-eq kg⁻¹ fish total GWP, corresponding to a 10% reduction relative to conventional composting. When burial treatments were compared, the soil-only system presented slightly lower GWP than the sawdust + soil treatment, avoiding 24 g CO₂-eq kg⁻¹ fish, which corresponds to a 4% reduction. The contrast between management strategies was more pronounced when composting and burial pathways were compared. Conventional composting avoided 261 g CO₂-eq kg⁻¹ fish, corresponding to a 44% reduction relative to the sawdust + soil burial treatment. These results indicate that aerobic waste management strategies substantially reduce the climate impact associated with fish waste disposal compared with burial systems.

3.5. Compost Characteristics: Thermal Behavior, Nutrient Composition, and Reductions in TS and VS

The thermal stability of biochar is an essential parameter for understanding its resistance to aging. The TGA–DSC curves of isolated biochar (Figure 8a) showed intensified mass loss around 500 °C, stabilizing near 600 °C, which indicates a higher concentration of aromatic carbon. This pattern was also described by [62], who identified the 500–600 °C range as indicative of the stability of biochars from different sources. When evaluated together with composting residues (Figure 8c,e), mass loss occurred at lower temperatures, around 300 °C. This suggests that, despite the intrinsic stability of biochar, more labile constituents of the system influence the thermal profile, making the residual material gradually more stable throughout the process.

At 130 days, two peaks were observed in both composting treatments, but in the material with biochar (Figure 8c) the second peak was less intense, suggesting adsorption or stabilization of recalcitrant compounds. This effect is relevant, since even after 100 days, composting with biochar maintained higher temperatures (Figure 8e), indicating ongoing microbial activity. Under these conditions, the additive limited the emission of volatiles, reinforcing its stabilizing role. The remaining mass was also greater in composting with biochar, likely due to the increased recalcitrant fraction resulting from its combination with sawdust. This outcome is consistent with [63], who reported higher residual mass in TGA following lignin addition.

In the burial experiment (Figure 8b,d), TG curves indicated initial water loss up to 240 °C, followed by organic matter decomposition. In the residues buried with sawdust and soil (Figure 8b), exothermic peaks were detected at 355 °C (135 days), associated with the release of volatile compounds. Conversely, in the burial with soil only (Figure 8d), no apparent peaks were observed, suggesting the absence of retained gases. This result aligns with the temperature monitoring, which did not exceed 45 °C, indicating minimal microbial activity.

As expected, the residual mass was higher in the burial with sawdust and soil (approximately 31%) compared with soil alone (18%). This effect is related to the greater recalcitrance of sawdust, which is rich in lignin [64], favoring the accumulation of material at the end of the process.

The final compost exhibited higher concentrations of N, P, and K in the Bulk and BulkBioch treatments (

Table 3. Characterization of the compost: nutrients (N, P, and K), total solids (TS) and volatile solids (VS).

Treatment	N (g kg−1)	P (g kg−1)	K (g kg−1)	TS (%)	VS (%)
Bulk	16.5	22.3	4.2	61.25	63.37
BulkBioch	17.4	30.8	4.6	67.46	67.46
BulkS	7.1	4.9	2.5	77.27	32.22
S	9.4	4.5	2.9	80.50	21.95

). The findings of nutrient concentration in this study are partially aligned with previous research on composting fish waste in aerated static windrows [65], which observed a variation of 2.7 to 3.0% in N, 21.3 to 25.4 g P kg⁻¹ compost, and 7.8 to 11.8 g K kg⁻¹ compost. The lower N level in our study can be attributed to the initial composition of substrates and organic waste used, and experimental conditions. However, biochar demonstrated effectiveness in retaining N in composting, similarly with results observed in composting of cattle slaughterhouse waste [66] and in sewage sludge [67].

The burial treatments (S and BulkS) resulted in composts with markedly lower nutrient concentrations, higher TS values, and very low VS contents (Table 3). These results may reflect the suppressive effect of soil addition on microbial activity, as indirectly evidenced by the lower average temperatures recorded in these treatments (Figure 9). Exploratory analysis suggested a strong correlation between nutrient concentrations and composting temperature (Appendix A, Table A1 and Table A2), although this result should be interpreted with caution, as the correlation analysis was based on treatment-level summary values from only four non-replicated experimental units. Soil incorporation dilutes the organic matter fraction, thereby limiting the energy available for microbial metabolism and preventing the development of sustained thermophilic phases [68].

3.6. Plant Growth Response to Compost Application

Compost application rate significantly affected all evaluated parameters (p < 0.05), with the greatest responses observed at the highest doses. As no significant effect of compost type was detected, data from both composts were pooled, and a single regression curve was used to illustrate the dose–response relationship.

The highest compost application rate resulted in lettuce plants with approximately twice the total length of those grown in soil without compost. It also promoted substantial increases in shoot biomass and root development, as evidenced by the linear relationships observed for shoot weight, shoot length, root weight, and root length (Figure 10a–d). These findings highlight the importance of organic fertilization in enhancing crop growth and quality. The gradual mineralization of organic matter in compost ensures a sustained release of nutrients, thereby improving nutrient availability and uptake efficiency by plants.

Similar results were reported by [68], who evaluated poultry manure-based compost amended with biochar during lettuce cultivation. The authors observed increases of 71% in plant biomass, 77% in soil organic carbon, and approximately 90% in the availability of macronutrients (N, P, and K), along with improvements in plant morphology and canopy diameter.

4. Conclusions

This study shows that fish waste management strategies may influence greenhouse gas emissions, process dynamics, and nutrient recovery during degradation. In this pilot-scale trial, composting amended with 10% biochar resulted in lower CH₄ and N₂O emissions, earlier CH₄ stabilization, and greater nutrient retention in the final compost. Methane emissions were more pronounced during the early stages of degradation, whereas N₂O emissions tended to increase toward the later stages of the process. Burial treatments, in contrast, were associated with lower decomposition efficiency and higher N₂O-related climate impact. In addition, composts derived from fish waste proved to be effective organic fertilizers for lettuce cultivation, with plant growth responding positively to increasing application rates. Overall, these findings indicate that composting with biochar and structural bulking agents represents a more sustainable alternative for fish waste management, with potential applications in circular-economy strategies aimed at waste valorization and greenhouse gas mitigation.