Introduction

The consumption of dietary animal protein has rapidly increased due to urbanization and rising disposable income in China. This trend is also observed globally, as socioeconomic development often leads to a shift from plant-based to animal-based food1,2. As a result, the global livestock and poultry industry has expanded, leading to the production of large amounts of excreta1,2,3. Wang et al.4 estimated that the total weight of manure produced in China in 2017 is 1.99 × 1012 kg, which is close to the 1.90 × 1012 kg reported by the Ministry of Agriculture and Rural Affairs of the PRC in 2015. Although these wastes can serve as valuable materials for anaerobic fermentation to produce biogas or fertilizers3,4they can also pose serious environmental risks if not properly managed. Discharging them into the environment without appropriate treatment can result in soil and water pollution3. Therefore, timely and effective treatment is crucial to enable the reuse of organic matter in these wastes and prevent environmental contamination5.

Livestock manure is often mixed with large volumes of urine and wash water, resulting in a highly diluted slurry. Therefore, solid-liquid separation is typically carried out before further processing of the manure excreta6,7,8. This makes the treatment of the liquid portion more practical and affordable6. Numerous studies have been conducted on the environmentally sound treatment and fertility of the solid and liquid fractions after separation1,9,10,11,12. For instance, constructed wetlands have been used to remove chemical and biological pollutants from swine wastewater, with the goal of reclaiming water for reuse13. Additionally, a self-sustaining synergetic microalgae-bacteria symbiosis system has been developed to achieve environmental sustainability12. The application of beetroot (Beta vulgaris) at a rate of 80 t/ha has also been shown to increase plant growth, marketable yield, and quality14. Furthermore, cow dung microorganisms have been found to naturally increase soil fertility through phosphate solubilization, and can also produce novel chemicals15. However, the direct use of raw livestock manure as fertilizer may pose a threat to crops due to the presence of insect eggs and pathogenic bacteria. To mitigate these risks, livestock manure is typically fermented before being used as a crop fertilizer4,16,17.

Fermenting the liquid fraction of livestock manure is a sustainable strategy for crop fertilization as it eliminates pathogens while retaining important nutrients such as nitrogen and phosphorus. Currently, various fermentation methods (FMs) have been studied18. One common practice in some European regions with intensive farming is anaerobic digestion, which produces a nutrient-rich digestate that can improve soil structure when applied in agriculture. This method is also efficient in reducing greenhouse gas emissions18,19. Additionally, membrane filtration is a technical solution that can effectively redistribute unbalanced nutrient concentrations in the liquid fraction20,21,22. However, due to economic and technical constraints, these technologies are not been widely used for treating the liquid fraction of livestock manure23. In China, traditional FMs are still the primary methods for managing the liquid fraction of livestock manure. However, their impact on fertilizing capacity has not been thoroughly evaluated. Although the germination percentage is an important indicator for assessing fertilizer availability24,25,26no prior studies have evaluated the fertility effect of fermented cow manure liquid fraction based on plant germination metrics. The purpose of this study is to assess the effects of different FMs on the potential of cow wastewater for soil fertilization, clarify the fertilizing characteristics of the fermented products, and provide reference for subsequent agricultural production and use. We compared the changes in pH, conductivity, total nitrogen (TN), nitrate, ammonia, water-soluble organic carbon (WSOC), total phosphorus (TP), and chemical oxygen demand (COD) of seven FMs, i.e., open (O), open with microbial agent (OMA), open with mixing (OM), shading ball (SB), G-membrane sealing (GS), H-membrane sealing (HS), and H-membrane sealing with microbial agent (HSMA), for 13 weeks. We also evaluated the potential of treated cow wastewater for fertilization by conducting germination experiments. These methods were selected because they are widely used in practice due to their low cost and operational simplicity, despite limited scientific evaluation of their effectiveness. This study provides fundamental data for assessing nutrient loss during fermentation of the liquid fraction of cow manure and supports the selection of optimal composting methods tailored to crop-specific nutritional needs.

Results

Changes in physicochemical parameters of liquid fraction of cow manure during fermentation

All seven FMs resulted in an increase in the pH of the liquid fraction of cow manure. However, the pH of the HS and HSMA methods stabilized earlier (by the fifth week after fermentation) (Fig. 1A). Conductivity exhibited clear variation among the methods. The HS method showed a steady increase over the first six weeks, followed by stabilization. In contrast, the HSMA method initially decreased during the first four weeks, returned to its baseline level by week eight, and then stabilized. Conductivity under the OM and GS methods slightly increased during the first four weeks and subsequently declined, ending at a level lower than the baseline by week 13. The O, OMA, and SB methods showed a consistent decrease in conductivity throughout the fermentation (Fig. 1B). The COD levels across all methods declined rapidly in the first six weeks and then plateaued (Fig. 1C). N levels in the HS and HSMA methods remained relatively unchanged, while other methods showed a gradual decrease, with the greatest reduction observed in the O, OMA, and SB methods (Fig. 1D). The WSOC decreased steadily in all methods, with similar magnitudes of reduction (Fig. 1E). The TP concentrations in the HS and HSMA methods continued to decline, whereas TP in the remaining methods decreased rapidly in the first five weeks and then stabilized (Fig. 1F). The NH4-N concentrations in the HS and HSMA methods remained almost unchanged, while the other methods decreased steadily (Fig. 1G). The NO3-N levels decreased consistently across all treatments (Fig. 1I). Germination percentage increased rapidly in the first five weeks, exceeding 80% in all treatments, and then remained stable (Fig. 1H). Although some fluctuations were observed, the SB and GS methods showed the most stable germination percentages above 90% after week five (Fig. S1). The trends in average root length and GI mirrored each other. After fermentation, the HS and HSMA treatments showed the lowest root length and GI, followed by the OM, while the other methods yielded higher values (Fig. 1J, K). Fecal coliform levels dropped below 20 MPN/L in all treatments by the fifth week (Fig. 1L).

Fig. 1
figure 1

Changes in physicochemical parameters during fermentation process. (A) pH; (B) conductivity; (C) chemical oxygen demand (COD); (D), total nitrogen (TN); (E) water-soluble organic carbon (WSOC); (F) total phosphorus (TP); (G) ammonia nitrogen; (H), germination percentage; (I) nitrate nitrogen; (J) average root length; (K) germination index (GI); (L) fecal coliform. Shaded areas are 95% confidence intervals. (O) Open; OMA, open with microbial agent; OM, open with mixing; SB, shading ball; GS, G-membrane sealing; HS, H-membrane sealing; HSMA, H-membrane sealing with microbial agent.

Full size image

Weekly conversion rates of COD, WSOC, TN, NO3-N, NH4-N, and TP revealed significant differences across treatments and fermentation stages (Kruskal-Wallis rank sum test with Dunn post-hoc test, P < 0.05; Fig. S2). For instance, in week 1, the HSMA treatment caused the fastest reductions in COD, WSOC, NO3-N, and NH4-N, while TN increased. The SB treatment produced the fastest TP decline (> 50%). Similar weekly variations were observed across the 13 weeks, with no consistent pattern indicating that microbial addition uniformly influenced nutrient changes. Furthermore, mixing (as in OM) was expected to enhance NH4-N oxidation by increasing DO levels, yet our results showed no significant reduction in NH4-N in open liquid compost with mixing (Kruskal-Wallis rank sum test with Dunn post-hoc test, P > 0.05; Fig. 2 and S2). Although different FMs significantly influenced COD levels (P < 0.05), post-hoc comparisons revealed that significant differences only occurred during the first nine weeks, with no significant differences observed after week 10 (Dunn post-hoc test, P > 0.05; Table S1).

Fig. 2
figure 2

Correlation between physicochemical parameters of liquid fraction of cow manure and germination. (A) Spearman correlation between physicochemical parameters of liquid fraction of cow manure and germination and fecal coliform; (B) correlation between water-soluble organic carbon (WSOC) and germination percentage (GP); (C) correlation between water-soluble organic carbon (WSOC) and germination index (GI); (D) correlation between NO3-N and GP; (E) correlation between NO3-N and GI; (F) correlation between TP and GP; (G) correlation between TP and GI; (H) correlation between NH4-N and GP; (I) correlation between NH4-N and GI; (J) correlation between TN and GP; (K) correlation between TN and GI; (L) correlation between COD and GP; (M) correlation between COD and GI; (N) correlation between conductivity and GP; (O) correlation between conductivity and GI; (P) correlation between pH and GP; (Q) correlation between pH and GI. The gray areas indicate the 95% confidence interval. *** p < 0.001.

Full size image

The GS method showed the highest TN concentrations in weeks 2 and 3, while the HS method maintained the highest TN from week 5 onward (Table S2). The NO3-N concentrations in GS were significantly higher than those in HSMA between weeks 3 and 6, and higher than in HS at week 13 (Kruskal-Wallis rank sum test with Dunn post-hoc test, P < 0.05; Table S3). For NH4-N, OM had significantly higher levels than OMA in week 1 and higher than O in week 2. HS had higher concentrations than SB in week 3 and higher than O and OMA in weeks 4–13 (Kruskal-Wallis rank sum test with Dunn post-hoc test, P < 0.05; Table S4). The TP concentrations varied over time. HSMA had the highest TP in most weeks, whereas SB had the lowest from weeks 2–8, OMA had the lowest in weeks 9–10, and OM had the lowest in weeks 11–13 (Table S5).

Effects of different FMs on the germination

Although FMs caused significant differences in germination percentage at week 13 (Kruskal-Wallis rank sum test, P < 0.05), Dunn’s test did not show any significant differences throughout the process (P > 0.05; Table S6). However, SB consistently produced longer root lengths, while HS yielded the shortest (Table S7).

Correlations between physicochemical parameters and germination traits

Spearman correlation analysis indicated that the germination percentage, GI, and root length were positively correlated with pH (r > 0.6, P < 0.05), and negatively correlated with WSOC, NO3-N, NH4-N, TN, TP, COD, and conductivity (r < − 0.6, P < 0.05). Regression analysis revealed that WSOC, NO3-N, TP, TN, COD, and pH were significantly associated with germination percentage in multiple linear models, while NH4-N and conductivity showed significant univariate associations (P < 0.05; Fig. 2). Similar trends were observed for GI and root length (Fig. S3). Fecal coliform counts showed positive correlations with WSOC, NO3-N, NH4-N, TN, TP, COD, and conductivity (P < 0.05), and negative correlations with pH (P < 0.05; Fig. 2).

Discussion

Cow manure is a cheap and readily available bioresource on our planet15,27. Currently, it is commonly used either for fermentation to produce renewable energy such as biogas28,29or as an agricultural organic fertilizer19. Organic manure derived from animal waste holds great promise, and has a clear advantage over chemical fertilizers in various aspects, including improving soil health and fertility, maintaining soil biodiversity, and enhancing crop productivity30,31,32,33. Gao et al.34 reported that in a 38-year fertilization experiments in double-cropping rice systems, he combination of N, P, and K with cow manure significantly improved rice yield and soil fertility, although balanced fertilization was necessary. Cheng et al.35 found cow manure compost increased the abundance of pollutant-degrading bacteria in black soil, but had inhibitory effects in saline-alkali soils. Moreover, while compost addition increased imazethapyr degradation by 12.6% in black soil, it decreased it by 7% in saline-alkali soils. Despite growing interest in manure as a fertilizer5,36,37,38the effect of different FMs on the fertilizing potential of the liquid fraction of cow manure has not been comprehensively evaluated. Most previous studies focus on reducing nutrient loads to avoid environmental pollution8with few addressing fertilizer effectiveness39. Our results indicated that different FMs significantly affected average root length and GI, although all methods raised germination percentage above 80% after six weeks. According to the Agricultural Industry Standard of China (NY/T 525–2021), a GI ≥ 70% is required. Only the O, OMA, SB, and GS methods met this standard after 13 weeks of fermentation. The lower GI in OM, HS, and HSMA methods may be due to their higher TN and NH4-N concentrations (Fig. 1), which were negatively correlated with germination percentage, root length, and GI (Fig. 2A). Therefore, diluting these fermented liquids or combining them with other fertilizers may help achieve acceptable GI values40.

A previous study reported that the fermentative process of mixed livestock manure reduced the pH, electric conductivity, TN, NH4-N, and TP, while increasing NO3-N41. In contrast, our results showed increased pH and decreased steadily in NO3-N, COD, WSOC, and TP during fermentation. This discrepancy likely arises from differences in FMs and conditions. Closed fermentation systems (HS and HSMA) retained high TN and NH4-N, whereas open systems (O, OMA, and SB) more effectively reduced these nutrients - except for OM. The H-membrane may restrict oxygen entry, limiting aerobic denitrification and reducing nitrogen volatilization. In the OM method, mixing likely released organic nitrogen into suspension but did not allow sufficient oxidation, resulting in high TN and NH4-N concentrations. Additionally, the addition of microbial agents had no substantial impact on fermentation outcomes. Nitrogen and phosphorus are crucial for fertilizer value42,43,44. Compared to chemical fertilizers, organic fertilizers release nitrogen more slowly, leading to higher nitrogen use efficiency, crop yield, and grain quality45,46. Previous studies showed that cattle and sheep manure composts had higher pH, NO3-N, total carbon, and C/N ratios, while pig and poultry composts had higher EC, NH4-N, TN, and TP47. Our results showed that all methods, except HS and HSMA, significantly reduced TN and NH4-N. In all treatments, NO3-N and TP declined steadily. Moreover, NH4-N levels in open fermentation were consistently lower than in closed systems during mid and late fermentation, indicating that aerobic conditions promoted ammonia oxidation. However, mixing had no substantial effect on NH4-N levels. Importantly, TN, NH4-N, NO3-N, and TP were negatively correlated with germination percentage, root length, and GI, supporting the recommendation to dilute fermented liquid manure or blend it with other fertilizers to reduce phytotoxicity.

Electrical conductivity (EC) reflects soluble salt content, which may negatively impact plant growth if too high40. Patel and Lakdawala48 noted that EC > 2.5 dS/m is toxic to most crops. All fermented liquids in our study exceeded this threshold, indicating that dilution or blending is necessary before application.

According to Patel and Lakdawala48optimal pH for liquid fertilizers lies between 6.5 and 7.8. Only HS and HSMA methods produced liquid with pH values within this range; the others exceeded 7.8. Although higher pH was positively correlated with germination percentage, further research is required to understand its effect on different crop species.

The fertilizer value of fermented manure depends on both crop species and application rate49. Valentinuzzi et al.49 reported that the liquid digestate releases nutrients faster than solid manure pellets. Our study evaluated seven FMs using cucumber seeds under laboratory conditions. Future work should assess the effects on other plant species and in field conditions.

Conclusions

Different FMs of the liquid fraction of cow manure caused varying effects on average root length, and GI. However, all FMs increased the germination percentage to over 80% after six weeks of fermentation. The HS and HSMA methods maintained TN concentration and conductivity throughout fermentation and kept the pH within an appropriate range, while the other FMs steadily reduced the TN and NH4-N content in the liquid fraction. All FMs steadily reduced NO3-N and TP. The TN, NH4-N, NO3-N, and TP were negatively correlated with germination percentage, average root length, and GI, indicating that the fermented liquid fraction should be appropriately diluted or used in combination with other fertilizers to mitigate the adverse effects of nitrogen and phosphorus on seed germination. These findings provide fundamental data for evaluating the impact of different FMs on nutrient loss in cow manure liquid fraction, which may assist in selecting the optimal composting methods based on crop-specific nutritional requirements. Moreover, this study is the first to assess the influence of various FMs on the humification of cow liquid manure and to propose seed GI as a potential indicator for evaluating humification. However, due to the limitations in experimental conditions, microbial composition and biological parameters such as biomass, enzyme activity, and substrate conversion rates were not assessed. Since the microbial community plays a key role in fermentation, future studies should include these biological indicators to clarify the mechanisms underlying the physicochemical changes observed across different FMs.

Materials and methods

Experimental site conditions and pretreatment of cow excreta

The liquid fraction of cow manure excreta was obtained from the Qinglong Base of Guangzhou Huamei Milk Co., Ltd. (113.391 E, 23.689 N; Guangzhou, Guangdong, China). The base had a stock of approximately 3,500 dairy cows, with a total of 8 breeding pens, was collected using scrapers and transferred to a cesspool. Well-mixed raw manure was then pumped from the homogenization cesspool to a solid-liquid separator. After dewatering, the liquid portion entered a dedicated excrement pool for treatment and subsequent discharge (Fig. 3). The wastewater used in this experiment was the liquid fraction obtained from this excrement pool.

Fig. 3
figure 3

Diagram of solid-liquid separation and on-farm photos.

Full size image

Experimental design and sample collection

The study was conducted at the outdoor experimental area of the College of Natural Resources and Environment, South China Agricultural University between October 8, 2021, and January 3, 2022. The average temperatures for each month during the experiment were 26, 22, 16, and 15 °C. The study consisted of seven common FMs, i.e., O, OMA, OM, SB, GS, HS, and HSMA, respectively (Table 1). Each method included three replicate treatment units, all of which injected 350 L liquid fraction for experiment into a 400 L white polyethylene drum with 75 cm of the bottom diameter, 93 cm of the top diameter, and 77 cm of the height (Longfei, Chnagzhou, Jiangsu, China; Fig. S4). G-membrane was a customized waterproof and ammonia permeable membrane. H-film was a HDPE anti-seepage film purchased from the market. The microbial agent used was Wangnongbao manure fermentation starter (Zhengzhou, Henan, China), with 360 mL of the mixed starter added to each treatment that included a microbial agent. In the OM method, the fermentation liquid was steadily stirred with an electric mixer at 50 rpm.

Table 1 Seven fermentation methods used in this study.
Full size table

The study lasted for 13 weeks, at weekly sampling on Monday. Considering the evaporation of the units, the units were filled with water to approximately 350 L and stirred manually before each sampling, and then 2-L samples were collected by the multi-point sampling method and transported to the laboratory at low temperature in an insulated box with ice bags for subsequent analysis.

Determination of physicochemical parameters

Conductivity and pH were measured using the HQ40d portable digital multi-parameter analyzer (Hach, USA) during sampling. The sample and surrounding temperatures were detected using a wireless automatic temperature recorder (Huahanwei, Shenzhen, China). TN was determined using ultraviolet spectrophotometry with alkaline potassium persulfate digestion50. Nitrate and ammonium nitrogen were determined using flow injection analyzer spectrophotometry as previously described51,52. WSOC was determined using the purple red complex spectrophotometry as previously described53. Total phosphorus was detected using molybdenum blue method54,55. COD was detected using spectrophotometry with COD digester (DRB200, Hach, USA)56. GI was detected using Baimeng Jinyan No. 4 cucumber seeds (PiShu Agricultural Technology Co., Ltd, Xuzhou, Jiangsu, China) according to a previous description57. Briefly, the number of germinated seeds was counted from the first to seventh days during the seed germination testing. Seeds were germinated when the radicle had emerged by more than 2 mm. The GI was calculated by:

$$\:\text{G}\text{I}\:=\:\sum\:\frac{{\text{G}}_{\text{t}}}{{\text{D}}_{\text{t}}}$$

where Gt is the number of germinated seeds on the tth day and Dt is the number of corresponding germination days. The average of three replicates was used for the data analysis.

Weekly conversion rates (CR) of COD, WSOC, TN, NO3-N, NH4-N, and TP were calculated according to the formula:

$$\:{CR}_{i}=\frac{{C}_{i+1}-{C}_{i}}{{C}_{i}}$$

where CRi is the ith conversion rate. Ci and Ci+1 is the concentrations of physicochemical parameter in the i and i + 1 weeks, respectively.

Data analysis

Data were enunciated as the mean ± standard deviation. Kruskal-Wallis rank sum test with Dunn post-hoc test was conducted using R FSA version 0.9.6 package (https://cran.r-project.org/web/packages/FSA/index.html) to test the differences in data between FMs. Spearman analyses were conducted using R psych version 2.5.3 (https://cran.r-project.org/web/packages/psych/index.html), reshape2 version 1.4.4 (https://cran.r-project.org/web/packages/reshape2/index.html), and corrplot version 0.95 (https://cran.r-project.org/web/packages/corrplot/index.html) packages. Linear correlation analysis was conducted using R basicTrendline (https://github.com/PhDMeiwp/basicTrendline) package.