Optimization of the process conditions for methane yield from co-digestion of mixed vegetable residues and pig manure using response surface methodology

To determine optimized conditions for co-digestion for a mixture of four kinds of mixed vegetable crop residues consisting of cucumber, tomato, eggplant, and pepper mixed in equal parts on a mass basis, co-digestion experiments were carried out with pig manure. The interaction effects of parameters such as manure-to-mixed vegetable residues ratios (M/S), initial pH, and organic load (OL) were investigated with respect to cumulative methane yield using response surface methodology (RSM). The highest cumulative methane yield was calculated to be 380.50 mL/g VS at an initial pH of 7.3, OL of 18.8 g VS/L, and M/S of 3.9:1. Comparison and veri�cation experiments showed that under optimized conditions the co-digestion process showed increased the methane yield and had practical application value. The microbial analysis showed that the relative abundances of bacterial taxa, such as Clostridium_sensu_stricto_1, Fastidiosipila, and Terrisporobacter, were all highest in the co-digestion samples under optimized process conditions (PV). Different types of methanogenic archaea taxa in PV samples were richer than other samples, which showed higher relative abundances of Methanogenium, Methanobrevibacter, Methanoplanus, Methanospirillum, and Methanobrevibacter. Thus, the co-digestion system of a mixture of vegetable residues and pig manure can enrich different types of methanogenic archaea taxa, which leads to increased digestion performance, and may strengthen process stability. Importantly, pig manure, mixed vegetable residues can be included into anaerobic digestion applications through co-digestion, thus enabling valorization of these substantial residues and can be engineered for applications.


Introduction
As a large agricultural country, China's annual production of crop residues is more than 900 million tons, with vegetable crop residues making up approximately 25.6% of the total [1].Vegetable crop residues are characterized by high moisture content, short storage time, and logistical challenges for transport di culty.Improper management can result in accumulation of large amounts of such mixed vegetable waste which leads to decomposition, giving rise to nuisances such as malodor, insects, while posing a pathogen risk, and overall represents a waste of resources and increased environmental pollution [1].
Moreover, vegetable residues are very diverse, and sorting is time-consuming, and labor-intensive, and it would be more e cient to directly digest mixed vegetable crop residues, but study is necessary.Pig manure is an important organic waste, of which the yield has increased signi cantly with the rapid development of modern animal husbandry.Approximately 4 billion tons of livestock manure are produced in China annually, and pig manure accounts for 68.1% of this amount according to the data released from the National Statistics Bureau [2].Pig manure has a high chemical oxygen demand (COD) and high concentrations of nitrogen, phosphorus, and fecal coliforms, which can profoundly threaten the environment and human health if left untreated [3].
Anaerobic digestion (AD) converts agricultural and livestock waste, such as crop residues and livestock manure, into methane and is an e cient treatment technology with sound economic and environmental bene ts.AD includes both mono-digestion and co-digestion, and while AD of a single feedstock (i.e., mono-digestion) represents a well-established mature technology, some de ciencies such as improper carbon/nitrogen (C/N) ratios, low biogas yields, and unstable operating performance can be still be addressed to improve performance [4].Anaerobic co-digestion (co-AD) is the simultaneous AD of two or more substrates, which can adjust the C/N ratio, enhance biogas yields, and dilute potentially problematic compounds (i.e., inhibitors), and improve the stability of the AD process, and often has better performance than mono-digestion [5][6].Co-AD of crop residues and livestock manure is now widely recognized by researchers as an attractive approach to AD [7].Zhang et al. (2021b) [3] investigated the effect of mixing ratio on co-AD of swine manure and rice straw, and the results showed that the highest methane yield (188.79 mL/g VS added ) was obtained at a 1:1 ratio, which was improved performance by 178.8% and 18.9% when compared to mono-digestion of swine manure or rice straw by, respectively.Shen et al. (2019) [8] subjected durian shell to co-AD with chicken manure, cow dung, and pig manure at different ratios and found that signi cant synergistic effects were observed in the co-digestion of durian shell and pig manure at 1:1 and 1:3 ratios, respectively.However, co-AD e ciency is not only related to the ratio of the mixture (M/S) but also closely related parameters, such as organic loading (OL) and initial pH [9][10].
Response surface methodology (RSM) is a mathematical statistical method widely used to optimize variables and obtain desired response values [11].RSM allows the determination of optimal experimental conditions, the correlation between individual variables and response values, and the interactions between individual parameters, and it is now widely used in the eld of AD [12].Using RSM, in combination with pig manure co-AD optimization experiments were carried out using mixed vegetable crop residues consisting of cucumber, tomato, eggplant, and pepper residues, which were mixed in equal parts on the basis of mass.The interaction effects of M/S, initial pH, and OL on the co-AD cumulative methane yield (CMY) were investigated using RSM to optimize the process conditions for the co-AD of mixed vegetable residues and pig manure, and key associated microorganisms were analyzed to provide a more in-depth understanding to foster more e cient utilization of these signi cant agricultural byproducts.

Substrates and inoculum
Crop residues were obtained from the Horticultural Innovation Base of the Academy of Agriculture and Forestry Sciences at Qinghai University.After air-drying under natural conditions, residues were crushed using an AQ-180E mill (Cixi City Nai 'ou Electric Appliance Co., Ltd, China) and passed through an 8-mesh nylon sieve (particle size of < 2.5 mm).Four kinds of residues, namely cucumber, tomato, eggplant, and pepper, were mixed in equal parts on the basis of mass, then processed at 25°C for 24 h with 0.6 mol/L KOH (moisture content was set at 70% (m/m)), and set aside.Pig manure was obtained from Qinghai Yufu Animal Husbandry Development Co., Ltd.The inoculated sludge was obtained from a stable agricultural biogas digester from Qinghai Agricultural, LLC operated with cow dung as the raw material.
The inoculated sludge was anaerobically preincubated for 7 days to minimize endogenous gas production.The feedstock characteristics are shown in Table 1.

Factorial design
A three-level-three-factor experiment was designed using the Box-Behnken design (BBD), and the experimental factor coding and levels are shown in Table 2. Three factors, initial pH (7.0, 7.5, and 8.0), OL (10, 20, and 30 g VS/L), and M/S (2:1, 3:1, and 4:1), were used as independent variables.CMY was used as the response value.

AD experiment
Methane potential testing was carried out using a fully automated methane potential tester (MultiTalent 203, Bipuhuarui Environmental Technology Co., Ltd, Beijing, China).A total of 17 runs were set up, as shown in Table 2, and co-AD was carried out according to the inoculum-to-material ratio of 2:1 (based on VS).The additions of each raw material are shown in Table 3.After the RSM experiments were completed, the mono-digestion of mixed vegetable residues (denoted as V), the mono-digestion of pig manure (denoted as P), and the co-AD of mixed vegetable residues with pig manure (denoted as PV) were subjected to comparison and validation experiments under the optimized conditions.The pure inoculum was set as the blank control, denoted as CK.The experimental period was 45 days.Three parallel experiments were performed for each group.

High-throughput sequencing
Samples were taken from day 5 to 40 of V, P, and PV, respectively, and blank control samples (CK) were taken simultaneously, totaling seven samples for microbiological analysis.Samples in each group were identi ed by the group and the digestion time (e.g., for the V group: V_5,V_40).Sludge sample genomic DNA was extracted using a Qiagen QIAamp Fast DNA Stool Mini Kit and quality was evaluated with 1% agarose gel electrophoresis, wherein a clear primary DNA band without signs of degradation was required for further analysis.DNA concentrations and purities were then evaluated using a NanoDrop 2000 ultra-micro spectrophotometer prior to sequencing, with requirements of sample concentrations > 10 ng/µL, total sample volume > 500 ng, and A 260/280 values ranging from 1.8 to 2.0.

Analytical methods and data analysis
The TS and VS contents of the prepared samples were measured in triplicate using standard methods [13].Sample pH was measured using a pH meter (pHS-2F, Shanghai INESA Scienti c Instrument Co., Ltd, Shanghai, China).Design-Expert 12 software was used for data analysis, and Origin 2018 software was used for graphing.

Methane yields
Figure 1 shows the daily methane yield (DMY) and CMY for runs 1-17.In experimental groups 1-17, the gas production started from day 1 of AD.The DMY decreased to a low point on day 3-4, which may be due to the hydrolytic acidi cation stage during this period, where the bacteria in the AD system play a role in decomposing organic matter.This led to the accumulation of volatile fatty acids (VFAs), inhibiting the activity of methanogenic archaea and reducing the methane yield.There were considerable differences in the DMY with different OLs.Two prominent peaks for DMY occurred with an OL of 10 g VS/L (runs 1, 2, 9, and 11), with the rst peak occurring on day 9 (28.), which may be due to the higher load and prolonged acidi cation time.Compared with the other OL conditions, the DMY at the end of AD still had a slight upward trend under the OL of 30 g VS/L, indicating that a considerable amount of organic matter remained undecomposed in the later stage under high OL conditions.When the initial pH was changed, the DMY trend differed only for the high OL (runs 3 and 4), with the DMY peak appearing earlier and with a comparatively lower value of 18.7 mL/g VS/d at the initial pH of 7 (run 3).
The DMY peak appeared later, with the more prominent peak of 24.41 mL/g VS at the initial pH of 8 (run 4).This phenomenon may be because the acidi cation in the AD system is more evident under higher OL conditions.Adjusting the initial pH can reduce the acidity of the system and enhance the DMY.The effect of different M/S on the DMY peak was substantial, and the DMY peak could be signi cantly increased when the M/S was large, which might be due to the higher methane yield potential of pig manure [14].
Upon 45 days of AD, the CMY of runs 1-17 showed a trend of rst increasing and then gradually stabilizing.Under the same AD conditions, the CMY decreased with an increase in the initial pH and increased with an increase in the M/S, where the highest CMY was 384.0 mL/g VS and the lowest was 313.7 mL/g VS.

Response surface regression model and analysis of variance
The Box-Behnken experimental design and its results are shown in Tables 2 and 4. The CMY from the co-AD of mixed vegetable residues and pig manure was set as the response value, and different levels of coded values of initial pH (A), OL (B), and M/S (C) were used as independent variables to establish the tted equations.The quadratic regression equation used in Design-Expert 12 software was as follows: As shown in Table 4, the model F-value was 13.46 with P < 0.01, which implies that the entire model regression was signi cant.R 2 was 0.9454, which indicates that this model explains 94.54% of the variation in the response values and the lack of the t term (P = 0.0685 > 0.05).These were not signi cantly different, indicating that the model t was good.The adequate precision of 12.7131 suggests that the model can be used to navigate the design space.This regression model and equation can be used to analyze and predict CMY from co-AD of mixed vegetable residues and pig manure.
From the p-values of each factor shown in Table 4, it can be observed that initial pH and M/S had a signi cant linear effect on CMY.In contrast, the linear effect of OL was not signi cant.The interaction effects of the three factors were not signi cant.The quadratic effects of initial pH and OL were signi cant, while the quadratic effect of M/S was not signi cant.The absolute value of the coe cients of each factor in the regression model re ects the strength of the in uence of the factor on the response value.According to Eq. ( 1), the absolute value of the coe cients of the three in uencing factors is A > C > B, which indicates that the in uence of each factor on the CMY is in the following order: initial pH > M/S > OL.Response surface optimization analysis The RSM was used determine optimized parameters for CMY from co-AD of mixed vegetable residues and pig manure.The response surface and contour plots are plotted (Fig. 2-4).Figure 2a shows that the slope of the response surface is larger between OL and the initial pH, indicating that OL and the initial pH had a higher interaction effect for co-AD.When the M/S was held constant, the CMY showed an increasing and then decreasing trend with increases in both initial pH and OL, where the initial pH had a more signi cant effect on CMY.Moreover, the magnitude of change in CMY after the change in OL was relatively small.The contour plots show that the CMY from co-AD was higher when the initial pH was between 7.2 and 7.4 and when the OL was between 15 and 25 g VS/L.Meanwhile, the CMY was low when both the initial pH and OL levels were high (Fig. 2b).
Figure 3 shows the response surface plots and contour plots between OL and M/S, with more shallow slopes observed in comparison to Fig. 2.This indicates a smaller correlation between OL and M/S.When the initial pH was held constant, CMY gradually increased with increasing M/S.CMY increased and then decreased as the OL was changed, but the variation was slight (Fig. 3a).As shown in Fig. 3b, CMY was highest when the OL was approximately 20 g VS/L, and the M/S was in the range of 3.5:1-4:1.When the M/S was low, the CMY was also relatively low.
As shown in Fig. 4, the response surface plots between initial pH and M/S differed from the previous gures.When the OL was held constant, the CMY gradually increased with an increase in the M/S at a low initial pH, and it was highest at M/S of 3.5:1 to 4:1.However, the CMY did not uctuate signi cantly with M/S when the initial pH was high.Figure 4b shows that CMY was the highest when the initial pH was 7.2-7.4,and when the initial pH increased from 7.6 to 8.0, the CMY decreased sharply.Thus, initial pH had a more signi cant effect on CMY from co-AD, which is consistent with the results of ANOVA shown in Table 4.
The optimal levels of the independent variables with desirable responses were established by numerical optimization using the Design Expert software program based on experimental results and statistical analysis by superimposing different surface responses.In this study, the optimal methane yield conditions for co-AD of mixed vegetable residues and pig manure were an initial pH of 7.3, OL of 18.8 g VS/L, and M/S of 3.9:1, which predicted a methane yield of 385.0 mL/g VS.Initial pH, M/S, and OL are the most common fermentation parameters in AD experiments [15][16][17].In this study, the overall degree of in uence of initial pH, M/S, and OL on CMY was in the following order: initial pH > M/S > OL, which was consistent with the RSM results.Among them, CMY of co-AD decreased with an increase in initial pH, and CMY was larger with an initial pH between 7.0 and 7.5.This suggests that pH can affect the activity of speci c anaerobic microorganisms, which in turn affects the methane yield [18][19].Zhou et al. (2016) [19] used pig manure (total solids at 7.8%) as feedstock for AD and found that the AD performance strongly depended on pH, with gas production and methane content of 16.6 L and 51.8%, respectively, at pH 7.0, which was signi cantly higher than tests at pH 6.0 and 8.0.Similarly, Carotenuto et al. (2016) [20] found that methane concentration in biogas was highest at pH = 7.0 under mesophilic conditions.Zhai et al. (2015) [21] found that the optimal initial pH for the co-AD of kitchen waste with cow manure was 7.5 for optimal methane yield (179.8 mL/g VS).Rao et al. (2004) [22] reported an optimal methane yield at a pH of 7.0-7.2.Meanwhile, CMY in this study increased with an increase in the proportion of pig manure in the M/S, probably because pig manure is rich in nitrogen, and the resulting change in C:N ratio has a positive impact on methane yield.However, the lignin in mixed vegetable residues is di cult to convert into methane, and the higher the proportion of mixed vegetable residues, the lower the amount of methanisable material.In addition, the variation of OL had the lowest in uence on CMY in this study.The optimal OL was 18.8 g VS/L, which indicated that the experimentally selected range of OL was suitable for AD.While OL did not lead to reduced AD e ciency at lower loadings, higher loadings were observed to overload the system or even cause malfunctioning and other problems.Reported optimal OL values for AD for similar residues such as tobacco stalks, cotton stalks, and durian shells were 20.2 g VS/L, 25.6 g VS/L, and 20.4 g VS/L, respectively [16, [23][24], and thus are comparative to the results of this study.

Comparison and validation experiments
Figure 5 shows the DMY and CMY from the comparison and validation experiments.As shown in Fig. 5, the DMY trends of P, V, and PV were different.Among them, two pronounced DMY peaks appeared in P on days 12 and 22, respectively, and the rst DMY peak was the largest at 27.0 mL/g VS/d.The DMY trend uctuated considerably in V, and the DMY peaks appeared on days 3, 9, and 12, with a maximum of 13.6 mL/g VS/d, and then entered a relatively stable methane yield stage on days 15-22, DMY then slowly declined.PV only showed a prominent DMY peak on day 13, which was 25.3 mL/g VS/d, and after the DMY peak, like V, it entered a relatively stable methane yield stage on days 15-22.A small amount of gas was still produced until the end of AD.The CMYs of P, V, and PV showed a trend of increasing and then stabilizing at 385.6 mL/g VS, 235.6 mL/g VS, and 380.5 mL/g VS, respectively.The PV was very close to the predicted value (384.9 mL/g VS) with a relative error of less than 5%, which indicated that the response model was valid.In addition, the ANOVA results showed that the difference in CMY between PV and P was insigni cant (P > 0.05).Moreover, PV was enhanced by 61.5% compared to V, indicating that co-digestion facilitates inclusion of mixed vegetable residues alongside pig manure, allowing the valorization of this under-utilized feedstock without sacri cing process performance, while also potentially increasing process stability.
Of note, the CMY of PV in this study (380.5 mL/g VS) was signi cantly higher than the highest CMY in literature reports from co-AD of pig manure with vegetable wastes (244.3 mL/g VS), pig manure with rice straw (188.8 mL/g VS), pig manure with corn straw (220 mL/g VS), and pig manure with Pennisetum hybrid (299.0 mL/g VS) [3,[25][26][27].Zhu et al. (2022b) [15] used 4.0% KOH to treat eggplant residues to obtain a CMY of 159.0 mL/g VS.Cai et al. (2019) [28] predicted the methane production performances of different vegetable wastes and found that the methane production potentials of pepper stem and potato stem were 120.1 mL/g VS and 187.4 mL/g VS, respectively.The CMY of mixed vegetable residues under the optimized conditions in this study was 235.6 mL/g VS, which was much higher than that obtained from mono-digestion of other vegetable residues.This may be attributed to the more balanced nutrient elements of a variety of vegetable residues after mixing.The optimized conditions were more conducive to the stable operation of the AD system, improved the AD performance and enable the use of mixed vegetable residues for AD applications without sacri cing process performance through co-digestion.

Analysis of microbial diversity
As shown in Table 5, the OTUs, Chao1 values, and Shannon values of bacteria were higher than those of archaea.The values of each index were greater at the beginning of AD than at the end, indicating that the bacterial community had a higher abundance and diversity at the beginning of AD.Meanwhile, the PV samples had higher values for each index, indicating that mixing multiple raw materials could balance the nutrients in the AD system and increase the diversity of the bacterial community.In the archaea community, all indices were higher in PV samples than in P and V samples, indicating that the archaea in the co-AD system of multiple raw materials was more diverse.

Bacterial community structure
Using a relative abundance above 1% for the classi cation of dominant groups, bacteria belonged to 12 phyla and 22 genera (Fig. 6).As shown in Fig. 6a, at the phylum level, Bacteroidetes (relative abundance of 27.09-35.15%)and Firmicutes (14.94-28.87%)were the most abundant phyla, with the former having higher relative abundance in P and the latter having higher abundance in V and PV samples.The difference in abundance between the two may be due to the difference in raw materials.Among them, Bacteroidetes undergo protein hydrolysis, and pig manure, as a protein-rich feedstock, facilitates this enrichment [29].The relative abundance of Bacteroidetes was higher in the P samples.In contrast, Firmicutes was higher in samples with the addition of mixed vegetable residues, probably because Firmicutes are associated with the degradation of cellulose fractions [30].The relative abundance of Synergistota was higher in the samples with higher methane yield, ranging from 12.6-13.8%and 12.7-16.1% in P and PV samples, respectively.It was high in the PV samples, suggesting that it may be a functional bacterial taxon with an important role in methane production.Proteobacteria (with a relative abundance of 5.9-13.8%)was higher in the late stage of AD, indicating its involvement in the degradation of various lignocellulosic substances at the end stage [31].
As shown in Fig. 6b, among the 22 bacteria taxa at the genus level, DMER64 was the most dominant taxon in all the samples.Its relative abundance in all experimental groups (> 17.80%) was greater than that of CK (9.4%), which is capable of hydrolyzing a wide range of proteins into several small molecules, such as VFAs and NH 3 [32].Smithella was a sub-dominant taxon with an average relative abundance of 8.2%.Its relative abundance in the early AD stage was signi cantly lower than in the end, and its abundance in V samples was lower than that in PV samples during the same period.Smithella is a propionic acid mutualistic degrading bacteria susceptible to propionic acid, and its abundance decreases with increasing propionic acid concentration [33].Its lower relative abundance in the early AD stage than in the late stage may be due to the high concentration of propionic acid in the system during the hydrolytic acidi cation stage on day 5 of AD, leading to its reduction.Syner-01 was also found to have a high relative abundance in PV samples and could participate in reciprocal acetic acid and amino acid oxidation, resulting in the simultaneous facilitation of hydrolytic acidi cation of organic matter and methane production [34][35].In addition, Clostridium_sensu_stricto_1 (with a relative abundance of 2.0-3.7%),Fastidiosipila (1.2-2.8%), and Terrisporobacter (1.1-2.6%) were found more frequently in V and PV samples and were the highest in the PV samples.Among them, Clostridium_sensu_stricto_1 is a cellulose-degrading bacterium that degrades cellulose and hemicellulose [36].Fastidiosipila can convert complex macromolecules of organic matter into VFAs and CO 2 [37].Terrisporobacter can degrade various carbon sources, such as xylose and cellodisaccharide [38].The relative abundances of Clostridium_sensu_stricto_1 and Terrisporobacter were higher on day 40.Moreover, the abundance of Fastidiosipila was higher on day 5, suggesting that Fastidiosipila was primarily involved in transforming organic matter in the early AD stage of the samples with added mixed vegetable residues.
Clostridium_sensu_stricto_1 and Terrisporobacter were primarily involved in metabolizing macromolecular carbohydrates, such as hemicellulose and cellulose, in the later stage.Thus, adding different mixed vegetable residues for co-AD may enrich cellulose-degrading, hydrolyzing, and acidifying bacteria in the system, which participate in residue degradation.

Archaeal community structure
A total of 7 orders and 12 genera were obtained at the order and genus level of archaea (Fig. 7).At the order level, among the archaea taxa capable of producting methane, Methanosarcinales was the most dominant in all the samples, with relative abundance ranging from 38.0 to 45.1%.Methanomicrobiales were the sub-dominant groups with an average relative abundance of 15.8%.Moreover, Methanobacteriales (1.6-2.1%)only appeared in PV samples (Fig. 7a).
At the genus level, Methanothrix (28.5-40.2%)was the most dominant taxon in all the samples, which can only utilize acetate for growth and is an acetic acid-nutrient methanogenic archaea [39].This indicates that the archaea of the co-AD system mainly rely on acetic acid to produce CH 4 .Meanwhile, abundant hydrogenotrophic methanogens, including Methanoculleus, Methanoplanus, Methanospirillum, Methanocorpusculum, Methanogenium, and Methanobrevibacter, were also present in each sample [40][41].Among them, the relative abundance of Methanoplanus was higher in the samples supplemented with pig manure (P and PV), ranging from 4.9 to 6.5%.It was the highest in PV samples during the same period.Methanospirillum was also the highest in PV samples, with a relative abundance of 1.5-1.6%.Methanogenium was detected in the samples supplemented with mixed vegetable residues (V and PV) and was the highest in PV samples, ranging from 2.2 to 2.5%.Methanobrevibacter was also detected in PV samples, with relative abundances ranging from 1.3 to 1.8%.In addition, Methanosarcina had a high relative abundance of 7.8-9.5% in the PV samples.Methanosarcina is a multifunctional methanogenic archaea capable of utilizing acetic acid, H 2 /CO 2 , and methyl substrate to produce methane [42].This suggested that the system of PV could cause enrichment of different types of methanogenic archaea.Overall, the richest methanogenic archaea taxa were present in the PV samples.This suggests that the system of co-AD is rich in diverse, balanced nutrients and is suitable for the growth and metabolism of various types of methanogenic archaea.This can lead to the enrichment of different types of methanogenic archaea taxa.

Conclusion
In this study, an optimal process for methane production from co-AD of mixed vegetable residues and pig manure was obtained using RSM with an initial pH of 7.3, OL of 18.8 g VS/L, and M/S of 3.9:1.The CMY under these conditions was 380.5 mL/g VS, which was higher than the methane yield of many co-AD wastes with pig manure.The optimized co-AD process was effective at increasing methane production and enriching microbial communities and has value for practical applications.Importantly, this work shows that through co-digestion with pig manure, mixed vegetable residues can be included into anaerobic digestion applications through co-digestion, thus enabling valorization of these substantial residues, while mitigating uncontrolled greenhouse gas emissions and environmental degradation if left untreated.Future work for this research should now focus on scaling this approach to pilot and industrial applications.

Declarations Figures
Daily       Archaeal community structure in the mixed vegetable residues and pig manure digestate samples at the order (a) and genus (b) levels.
methane yield and cumulative methane yield for factorial design experiments comparing parameters relevant for co-digestion of mixed vegetable residues and pig manure.Open and closed symbols represent DMY and CMY respectively.

Figure 5 Comparison
Figure 5

Table 1
Physicochemical characteristics and pH of mixed vegetable residues, pig manure, and digester inoculum

Table 3
Added amount of each processed materials according to factorial design

Table 4
Analysis of variance (ANOVA) for the response surface quadratic model

Table 5
Microbial richness and diversity in samples from the co-digestion of mixed vegetable residues and pig manure