In order to achieve a sustainable agriculture, it is necessary to reduce the need for mineral fertilization input. The FAO Voluntary Guidelines on Sustainable Soil Management (VGSSM) encourage the adoption of agricultural practices that build and retain soil organic matter ( FAO, 2017 ). The total or partial replacement of mineral fertilizer with organic amendments is a good practice to solve the problems of the excessive use of mineral fertilizer while improving physical and chemical soil properties, carbon stocks and soil biodiversity ( Gattinger et al. , 2012 ; Sradnick et al. , 2013 ). The application of manure and compost on agricultural lands has shown a positive increase in water and nutrient retention, nutrient cycling, carbon transformation, soil biodiversity, soil structure and soil aggregation while enhancing soil organic matter content ( Treonis et al. , 2010 ; Nair & Ngouajio, 2012 ) and suppressing soil-borne pathogens ( Zaccardelli et al. , 2013 ). The supply of manure to agricultural soils is an ancient practice and a well-tested strategy to increase soil organic matter (SOM), replenish basic plant nutrients, improve yield response to fertilizers and restore soil productivity in degraded areas ( Schröder, 2005 ; Rufino et al. , 2007 ; Bogaard et al. , 2013 ; Nezomba et al. , 2015 ). At the same time, excessive manure applications and inefficient manure storage practices can have detrimental effects on the environment at multiple scales, such as the contamination of water and soil resources at local and regional levels, and the increase of greenhouse gases emissions (GHG) at a global level ( Sutton, 2011 ; Tubiello et al. , 2013 ). Manure is rich in nitrogen, phosphorus and carbon ( Pagliari & Laboski 2012 ) and can contribute to increase global climate change through the emission of methane and nitrous oxide ( Leytem et al. , 2011 ). Thus, while the recycling of livestock manure within agricultural systems is necessary to improve and maintain soil health, its efficient management is also important to reduce the environmental impact of farming activities. Therefore, manure treatment becomes a focal issue in relation to current national policies on environmental, climate and renewable energy matters.

Uruguayan agriculture has intensified during the last 15 years, with a large input of mineral fertilizer and other agrochemicals. In recent years, the application of manure has attracted increased attention, and nowa­days a group of Uruguayan farmers is incorporating a new technology based on an in situ aerobic process of the manure treatment that consists of continuous liquid composting (CLC) in aerated tanks. There are 52 tanks installed in the country, and the liquid manure composting is applied throughout ~ 8,000 ha. The assessment of soil quality after CLC application is essential to track changes in soils as a result of management practices. Within the framework of sus­tainable agricultural production, a high soil quality and health should maintain a high productivity without significant soil or environmental degradation ( Govaerts et al. , 2006 ; Griffiths et al. , 2010 ; Bünemann et al. , 2018 ). A set of biological soil indicators must be selected in order to advise farmers as to whether this management strategy yields the positive changes they anticipated.

Soil biological indicators are considered good soil quality indicators due to their sensitivity and ability to reflect soil management effects ( McGuire & Treseder, 2010 ; Bowles et al. , 2014 ; Benintende et al. , 2015 ). Generally, soil health has been related to the soil organic matter content ( Gao, 2006 ), soil fertility and structure, total N and C/N ratio ( Murphy et al. , 2011 ), but these soil properties do not necessarily change as a result of changing external conditions or use ( Chodak & Niklińska, 2010 ; Preem et al. , 2012 ; Muscolo et al. , 2014 ), and hardly reflect short-term changes in soil processes asso­ciated to new environmental threats. Changes in soil characteristics or environmental con­ditions induce rapid changes on microbial biomass, community composition and activity ( Schloter et al. , 2003 ; Gil-Sotres et al. , 2005 ; Nogueira et al. , 2006 ; Qin et al. , 2010 ). In some cases, changes in microbial communities and activity can precede detectable changes in soil physical and chemical properties, thereby providing an early sign of soil improvement or an early warning of soil degrada­tion ( Muscolo et al. , 2015 ).

However, there is still a lack of consensus on what to measure for the evaluation of soil biota linked to soil health and the ensuing prediction of sustainability or productivity ( Saha et al. , 2008 ; Dong et al. , 2014 ). Most of the studies about biological soil quality indicators propose measuring key functional microbial groups, as well as specific biological properties such as soil enzyme, microbial biomass carbon or N, ba­sal respiration, FAME profile, genomic analysis, and glomalin concentration ( Huang et al. , 2014 ; Dose et al. , 2015 ). These parameters can be used for the evaluation of soil health and as early indicators of soil degradation. However, the characterization of mul­tiple microbiological soil parameters to enhance the understanding of the correlation between soil biota and positive or negative effects on microbial functions and ecosystem services has not been as widely regarded.

The objectives of this study were: (i) to evaluate the influence of aerobically treated liquid manure (CLC) on soil microbial properties on two different agricultural sites with short- and medium-term CLC application (1 and 3 years, respectively), and (ii) to identify a minimal data set of microbiological parameters to monitor early changes induced by soil management with CLC as a biofertilizer.

Fifteen different microbial variables related to seve­ral soil biological functions and seven physicochemical parameters were analyzed (Table 2). The study showed variations in most of the evaluated microbiological properties of treated and untreated soil in each site (S1y and S3y, with one and three years of CLC appli­cation, respectively) (Table 3). S1y and S3y are adjacent sites with Eutric Cambisols . According to the physicochemical parameters analysed (Table 2), the two sites showed differences in SOC% and P content before the application of CLC. To avoid site effect, the main treatment for statistical comparison was presence (CLC) and absence (NCLC) of CLC application within each site. The statistical analysis of the main soil physicochemical properties within each site did not show changes with CLC application (Table 2). However, NMP was statistically higher ( p <0.05) in CLC amended plots at both sites.

MBC increased with the application of CLC in both sites. The increase was 13.0% in S3y and 23.3% in S1y; the increase was statistically significant only in the latter case (Table 3). The application of CLC had positive and negative significant effects on the mesophilic aerobic culturable microorganisms. Increases in cfu g ^-1 of MAB at S1y ( p <0.1) and MAA at S3y were recorded, as well as a decrease in MFF at S3y ( p <0.05, Table 3). These results indicate that the effects of the application of CLC on the recorded increases or decreases in the mesophilic culturable population were not consistent.

In both sites (S1y and S3y), there was a significant increase in the MPN of the functional groups as a result of the application of CLC (Table 3). In both sites, the application of CLC caused an increase in the values of cellulolytic and diazotrophic microorganisms. However, the CLC and NCLC treatments were only statistically different in S3y ( p <0.05). The density of P solubilizing bacteria was lower in CLC treatment in both sites, but the differences were only significant in S3y ( p <0.1). MPN of denitrifying microorganisms increased with CLC application in both sites (Table 3). In addition, we recorded significant positive correlations between MFF and P solubilizing microorganisms; MAB and NH ₄ oxidizers; MAA, N-fixing microorganisms and denitrificants (Table 4).

The activities of four enzymes (FDA, DHA, AlPh, AcPh) increased in both sites with the application of CLC (Table 3). However, the changes in FDA and DHA activities were not statistically significant. AcPh activity increased significantly in CLC treated plots at both sites (S1y, p <0.05; S3y, p >0.1).

In the case of glomalin, contrary to what was expected, our results did not show a significant change in GRSP with CLC treatment in neither site (Table 3). Glomalin was positively correlated with MAA and negatively correlated with enzyme activities ( p <0.05, Table 4).

SPI decreased with CLC treatment, although the differences were only significant in S3y ( p <0.05, Table 3). Therefore, a medium-term application of CLC (3 years) significantly increased the suppressing capacity of soil pathogens. Positive and negative significant correlations were found between SPI, enzymes and some functional groups (Table 4).

The results of the PCA carried out with phy­sicochemical and biological soil properties indicate that the two first components (PC1 and PC2) accounted for 88.3% of the total data variation (Fig. 1). The first axis opposed sites S1y and S3y (61.5% of variance explained). The biological soil properties DHA, AcPh and FDA, associated with S1y, and denitrifying, SPI, MAA and glomalin, associated with S3y, were important for the ordination of treatments in PC1. The second axis grouped CLC and NCLC treatments at each site (26.8% of variance explained). On the other hand, P solubilizers and MFF, associated with NCLC, and P, associated with CLC, were important for the definition of PC2. The results also indicated that the differences between the amended and non-amended soils were more pronounced in S3y than in S1y (Fig. 1). For this reason, and for the purpose of establishing the relationships between soil properties after three years of CLC application, another PCA was performed with the data from S3y (Fig. 2). The results indicate (Table 3) that the soil with three years of CLC application was characterized by higher values of N fixation, AcPh enzyme activity, actinobacteria and cellulolytic microorganisms, and lower values of P solubilization, MFF, oxidizing NH ₄ +, and SPI than NCLC.

This research provides insight into how microbial communities function and how composition responds to a short-term (S1y) and long-term (S3y) application of CLC. In accordance with other authors ( Giacometti et al. , 2014 ), long-term manure fertilization increases MBC. In this study, MBC was sensitive to CLC application and to the length of time of the application. The results obtained in this study showed that CLC application modifies some soil microbial properties and, in some variables, the effect becomes significant with time. MBC is considered an active component of the soil organic pool. Ren & Stefano (2000) suggested that Cmic:Corg ratios obtained over long-term treatments also represent C equilibrium in the soil. The results obtained at both sites indicate that CLC application increased MBC but not total soil SOC (Table 2). Thus, the changes in soil microbial biomass carbon measured over relatively short periods can indicate the trends in total organic matter content long before these can be detected by chemical analyses. Therefore, MBC as a biological parameter involved in soil C mineralization needs to be monitored together with SOC to predict SOC stability. Although many authors have reported positive correlations among the activities of soil enzymes, the microbial biomass and organic matter contents ( Wang et al. , 2011 ), it was not confirmed in this study. NMP increased significantly ( p <0.05) in both sites with CLC amendments (Table 2). Those differences were not likely due to different soil N content but rather to differences in microbial activity stimulated by CLC application ( Qiu et al. , 2008 ) warranting further investigation.

In this study, the abundance of functional groups was influenced by CLC amendments. Organic amendments, including manure application, modify the C/N ratio and the P demand for the development of microbial biomass ( Qiu et al. , 2008 ; Luo et al. , 2018 ). CLC application caused increases in NMP of N-fixing organisms and in acid phosphatase enzyme (AcPh) activity. On the other hand, P-solubilizing abundance decreased. Both variables, P-solubilizing and AcPh, are related to microbial activity and P availability, but linked to different soil P pools. P solubilizing was positively correlated with filamentus fungi (MFF) and AlPh to mesophilic aerobic bacteria (MAB) population. This may explain the different activity and origin of P availability in the soil. The majority of the variation in potential enzyme activity and functional groups could be explained by nutrient availability related to CLC application. In both sites, MBC was higher in CLC treatment than in NCLC, but differences were only significant at S1y ( p <0.1). Soil fertility is related to the activity of functional groups of microorganisms, with the ability to directly or indirectly supply essential plant nutrients because they are linked to N, P and C biogeochemical cycles. Trends in microbial biomass and functionality were not accompanied by a change in the size of the microbial culturable mesophilic population of bacteria, fungi and actinomyces (Table 3). The relationship between inorganic N and soil microbial activity, enzymes or functions is not straightforward because the number of microorganisms represents the number of culturable mesophilic microbial organism and not necessarily the active or functional groups in the soil. In general, microbial activity changes more quickly in response to management than to community composition ( Burger & Jackson, 2003 ). These results indicate that the number of mesophilic aerobic cul­turable organisms do not show consistent changes with CLC treatment, although quantitative changes of functional groups were recorded in both sites. Cellulolytic microorganisms break cellulose chains into small units and the treatment-enhanced activity of the cellulolytic community could indicate higher litter or root turnover in the soil and therefore a higher C pool. At the same time, the treatment also increases diazotrophic organisms with. An increase in microbial N demand due to CLC fertilization is expected. Furthermore, N ₂ O and N ₂ production is further stimulated by the addition of labile C sources typically added with organic-based fertilizers such as manure ( Zhou et al. , 2013 ). Increased N fixation could be a consequence of increased N ₂ emission with CLC application ( Asgedom et al. , 2014 ; Gao et al. , 2016 ); thus, microbial activity would regulate N input and output from the system. In this study, microbial community composition explained little variation compared to potential enzyme activity and functional groups. The phenotypic plasticity of the microbial community in response to treatment may be high, as suggested by the relatively large variation in soil microbial community functions. There were no statistically significant differences in soil pH among treated and untreated soils (Table 2). Many authors have suggested that manure amendments may decrease soil pH. It is known that soil pH affects microbial biomass and activity, as well as the relative proportions of bacteria and fungi ( Pietri & Brookes, 2008 , 2009 ). In this study, soil pH was positively correlated with enzyme activity and bacteria (MAB) and negatively correlated with actinobacteria (MAA) (Table 4). This positive correlation suggests that soil enzyme activity is mainly a function of pH and of the total present MAB population for potential synthesis ( Zhou et al. , 2019 ). The soil bacteria community abundance and composition were well adapted to soil pH for optimum enzyme activity ( Nannipieri et al. , 2017 ).

The suppressing capacity of soil pathogens (SPI) increased with CLC treatment in both sites. Acti­nobacteria (MAA) had a strong relationship with SPI, together with glomaline, diazotrophs and denitrifying organisms ( p <0.05, Table 4). Actinobacteria have gained special relevance as the most potent source of antibiotics ( Kandasamy et al. , 2012 ) and other bioactive secondary metabolites ( Solecka et al. , 2012 ) as potential biocontrol agents.

The activities of four enzymes (FDA, DHA, AlPh, AcPh) increased in both sites with the application of CLC. Soil enzyme activities are important indica­tors of microbiological and biochemical processes because they are involved in soil organic matter decomposition, nutrient cycling and availability, and in the biodegradation of toxic organic pollutants. Recent studies in temperate environments have suggested that measurements of soil enzyme activity are generally the most sensitive indicators of changes in the be­low-ground microbial community from different management practices ( Nanipieri et al. , 2002 ; Chaer et al. , 2009 ; Bowles et al. , 2014 ; Raiesi & Beheshti, 2014 ). Activities of specific enzymes may change depending on the composition of the organic amendments and the relative availability of nutrients, as well as other factors, such as soil type. Our results indicate that CLC application tends to increase soil enzyme activity and this pattern could be the result of higher microbial activity stimulated by CLC amendment. At the same time, given the relatively constrained level of soil nutrients, enzymatic activity might enhance the availability of the most limiting nutrients in order to meet microbial metabolic demands ( Sinsabaugh et al. , 2008 ; Mooshammer et al. , 2014 ), which could explain the significant increases of AcPh enzyme activity as a result of CLC application in both sites ( p <0.05, Table 3). In order to adapt to environmental constraints, microorganisms have to compensate the regulation of extracellular enzyme production with a C and nutrients acquisition strategy or by enhancing microbial metabolism for its nutrient use efficiency. In this study, the AcPh activity (linked to organic P), together with MPN of diazotrophic (N-fixation), celulolitic microorganism (C) and saprophytic fungi (MFF) increased significantly with CLC treatment. This could indicate a microbial response to P organic deficiency and changes of C/N ratio with treatment for the development of microbial biomass. These may partially help resolve the constraint of resource availability and explain the increase in enzyme ac­tivity. In this study, all enzyme activities increased; DHA showed a particular increase with CLC treatment in both sites. Among the various soil enzymes, DHA is of great relevance, as its activity levels are considered an indicator of overall microbial activity, due to their intracellular presence in all living microbial cells, and an indicator of biological redox systems.

Contrary to what was expected, our results showed no significant change in GRSP with treatment. Accor­ding to Zhang et al. (2014) , organic amendments enhance soil aggregate stability through the positive effect on soil binding agents, including GRSP. Recent results have described glomalin as a primary soil com­ponent in the formation of micro-aggregates and the maintenance of soil structure ( Wright & Anderson, 2000 ; Emran et al. , 2012 ). Results of Wang H et al. (2017) and Wang CY et al. (2017) showed that the enzyme activities after a 23-year manure amendment led to an increase in the macro-aggregates (> 1 mm) but not in the micro-aggregates. The latter could explain the negative correlation between glomalin and enzyme activity obtained in this study. Our results showed that CLC treatment increased enzyme activity and MBC but not total extractable glomalin. This may indicate that microbial biomass and enzyme activity are not linked to glomalin, which is specially related to micro-aggregate formation ( Six et al. , 2004 ). Instead, glomalin was positively correlated with MAA and SPI.

In this study, the sampling strategy and microbial parameters used revealed useful information regarding the effects of CLC application at the two sites, and this effect seems to be more evident with time (S1y compared to S3y). This is one of the first studies evaluating the effects of aerobically treated manure application on a large set of soil microbial parameters. The application of CLC had significant effects on biological soil properties, increasing microbial bio­mass, soil enzyme activity and the abundance of microorganisms of different functional groups. Among the analyzed microbial parameters, microbial biomass carbon (MBC) and soil enzyme activity of tested FDA, DHA, AcPh and SPI emerged as reasonable indicators to assess and monitor the effects of CLC application. Among physicochemical variables, N-mi­ne­ralization potential (NMP) was the best one to discriminate between treatments. The selected soil microbiological properties and NMP were the most sensitive to CLC amendments, but different soil types or sources of organic amendment applications need further research. With no current consensus on the minimal biological data set to assess the impact of a management practice on soil, this research approach provides insight into how microbial communities function and how composition responds to organic fertilization with CLC and its potential relationship with ecosystem services.

The authors would like to thank Andres Peres del Castillo and Christian Decker for their much-appreciated assistance with field work and Sally Bunning for her technical editing of the manuscript.