Agronomic and environmental performance of 6 energy crops in a close loop with digestate fertilization

ABSTRACT This study examines the growth performances, nutrient uptake and potential methane yield of six energy crops fertilized with the digestate liquid fraction (DLF) with or without arbuscular mycorrhizal fungi (AMF) inoculation. The 3-year-experiment involved the four perennial (Arundo donax L. Miscanthus × giganteus Greef et Deu, Heliantus tuberosus L. Lolium perenne L.) and two annual (Zea mays L. and Sorghum bicolor (L.) Moench × sudanense Stapf.) crops. The cumulative highest yield was observed for A. donax (128.1 ± 8.4 Mg ha−1), followed by M. × giganteus and S. bicolor (83.0 ± 8.0 Mg ha−1), H. tuberosus and Z. mays (66.2 ± 7.7 Mg ha−1) and L. perenne (31.4 ± 2.9 Mg ha−1). A. donax showed the highest estimated methane yield per hectare (11.030 Nm3 ha−1). After 3 years, the fertilization with DLF increased the initial soil Na+ content by 76.9%. During the experiment, only temporary and marginal effects were shown by AMF inoculation, which did not affect biomass production by any crop. Interestingly, AMF significantly increased the NO3-N concentration (+70%) while it reduced the NH4-N concentration in percolation water (−32.8%). DLF could be considered a viable organic fertilizer for biomass production, but the soil Na+ concentration should be carefully monitored.


Introduction
Bioenergy crops are gathering a lot of attention as the demand for innovative and alternative energy sources to fossil fuels is rapidly growing (Directive 2001/18/EC 2018).These crops are mainly perennials and are characterized by a high yield potential, rapid growth, and low input requirements (Rossini et al. 2019), even though increasing nitrogen (N) supply can trigger the biomass productivity of some of them (Roncucci et al. 2015;Ahmad et al. 2022).
The current framework of EU's Farm to Fork strategy calls on the agricultural sector to switch from fossil fuel dependency.In this context, the introduction of alternative fertilizers able to cope with increased fertilizer needs and the valorization of by-products of energy systems has been fostered.
Digestates represent a good alternative to synthetic fertilizers: as by-products of anaerobic digestion (AD) -a process able to efficiently convert low-value feedstock into biogas -they are a renewable form of fertilizer (Ronga et al. 2019).Moreover, due to their high organic matter (38.6-75.4% in dry matter) and nutrients (total N (TN) in dry matter: 3.1-14.0%,total phosphorus (TP) in dry matter: 0.2-3.5%,total NH 4 + in fresh matter: 0.15-0.68%)content (Nkoa 2014), digestates have been widely used as organic amendments or fertilizers with positive effects on the soil quality (Walsh et al. 2012).
Symbiotic mycorrhizal fungi such as arbuscular mycorrhizal fungi (AMF) are a significant component of the soil microbial populations and influence soil fertility, crop yield (Agathokleous et al. 2022;Langeroodia et al. 2022) and ecosystem sustainability (Gianinazzi et al. 2010).Symbiotic associations between AMF and plant roots can improve plant nutrition (Bhantana et al. 2021), water uptake (Augé et al. 2015), and resistance to soil-borne pests and diseases (Mitra et al. 2021).However, many agricultural practices including the use of fertilizers (in particular those with a high P content) and biocides, tillage, and monocultures are detrimental to AMF (Jach-Smith and Jackson 2018).
Several authors have studied the effect of the digestate liquid fraction (DLF) on soil microbial populations (Walsh et al. 2012;Bachmann et al. 2014), but little is known about its effect on plant mycorrhization.
The objectives of this study were i) to assess the valorization potential of digestate as fertilizer attainable with six energy crops in terms of biomass and energy production; ii) to verify if AMF inoculation can stimulate yield and nutrients (N and P) uptake iii) to gain indications on the potential environmental impact of digestate using the N and P concentrations in percolated water as a proxy.The experiment was laid out with a split-plot design, with AMF as the main plot and the crops as the sub-plots, replicated four times.Each replicate was represented by one growth box (2 × 2 m) installed with its top surface 1.3 m above ground level to avoid the influence of the water table, and its bottom open to allow water percolation.The boxes were filled with fulvi-calcaric Cambisol soil, according to the FAO-UNESCO classification (Table 1).A semi-automatic tension-controlled ceramic suction plate system was installed in two boxes for each treatment at a depth of 0.90 m to collect percolated water.The system was controlled by an electric vacuum pump activated by hand to maintain the ceramic plate suction at 0.02 Mbar.

Experimental setup
The DLF was applied once a year in spring at the dose of 250 kg N ha −1 in all boxes, according to the limit for the management of nitrate-vulnerable zones.The main chemical characteristics of the DLF are reported in Table 2. AMF (a granular inoculum of Rhizophagus intraradices, Funneliformis mosseae, Glomus etunicatum and G. clarum obtained from MycAgro Lab., France), selected among the most successful mix available on market, was inoculated during sowing or plant transplanting at a dose of 500 propagules m −2 , only in 2014 for perennial herbaceous crops, and in 2014 and 2015 for the annual ones.No AMF inoculation was carried out at the beginning of the 2016 crop season in annual crops to evaluate the persistence and success of the AMF inoculum from the previous two years.After adequate soil preparation, giant reed, miscanthus and Jerusalem artichoke were transplanted on 26 February 2014 at 7 rhizomes m −2 , perennial ryegrass was sown on 24 April 2014 using 5 g seeds m −2 .Sorghum was sown (4.5 g seeds m −2 ) on April every year as well as maize, which was thinned after emergence to obtain 7 plants m −2 .The boxes with maize and sorghum were interchanged annually.The experiment was performed under rainfed conditions.The growth boxes were kept free of weeds manually in spring every year for annual crops, but only in the first 2 years for perennial crops.

Meteorological variables
The main meteorological variables (rainfall, maximum, minimum and average temperature, wind speed and solar radiation) and reference evapotranspiration (ET0) were collected throughout the experimental period using a weather station located about 500 m from the experimental site (ARPAV, Legnaro).

Biomass sampling and analysis
Plant harvest was scheduled each year according to the literature about biomass harvested for methane (CH 4 ) production (Ragaglini et al. 2014;Dragoni et al. 2019).Giant reed, miscanthus, maize and Jerusalem artichocke were harvested once a year, sorghum was harvested three times a year, while perennial ryegrass was harvested five times in 3 years (Table S.1).At harvest time, fresh biomass production was measured in each growth box.A subsample of fresh biomass was used to determine the biomass moisture content by oven-drying at 65°C until constant weight.
The dried samples were ground to 2 mm (Cutting Mill SM 100 Comfort, Retsch, Germany) and used for tissue nutrient analyses.The N concentration was determined with the Kjeldahl method, while the P concentration was determined according to Balthrop et al. (2011).N and P uptakes were calculated as the product of the nutrient concentration and the aboveground dry biomass yield.

Root sampling and analysis
Root samples from three randomly selected plants per box were collected with a hand-operated soil probe (5 cm diameter) in the upper 20 cm of soil (30 cm for giant reed) in June of each growing season.The roots were washed, and all the soil particles were removed with a few drops of Tween 20 followed by rinsing with tap water, following Vierheilig et al. (1998).After that, the root samples were cleared with 10% KOH (45°C) for 1 h, rinsed three times with tap water, stained with 0.1% cotton blue in 80% lactic acid overnight, and then de-stained in 80% lactic acid for 48 h.
The percentages of AMF colonization were estimated according to Trouvelot et al. (1986), considering the following parameters: F%= mycorrhization frequency (the percentage of root fragments showing fungal colonization), M%= AMF colonization intensity (the percentage of fungal structures as referred to the whole root system), m%= AMF colonization intensity (the percentage of fungal structures as referred to the colonized root fragments), a%= abundance of arbuscules (the percentage of arbuscules as referred to the root fragments showing fungal colonization); A%= abundance of arbuscules (the percentage of arbuscules as referred to the whole root system).

Soil sampling and physico-chemical attributes
The volumetric soil water content was measured every 10 cm in the 0-90 cm soil profile with a Diviner 2000 device (Sentek, Stepney, Australia), which consists of a probe and a hand-held data logging display unit allowing in situ measurements.Data were collected weekly in the 2 years between the first and third DLF application (from April 2014 to April 2016).

Water sampling and analysis
A total of 252 percolation water samples were collected and analyzed for TN, NH 4 -N, NO 3 -N, TP and PO 4 -P.All samples were frozen immediately after collection and stored until laboratory analysis.TN and TP were determined using Valderrama's method (Valderrama 1981).PO 4 -P, NO 3 -N and NH 4 -N were determined using ion chromatography (Dionex ICS-900).

Methane yield of the crops
The CH 4 yield of the harvested biomass in normal cubic meter per hectare (Nm 3 ha −1 ) was estimated by multiplying the average aboveground dry biomass yield per crop throughout the experiment by a reference CH 4 potential in normal liter of methane per kilogram of volatile solids (NL CH 4 kg VS −1 ) for the same crop harvested with a similar percentage of dry matter (Table 3).

Statistical analysis
Data normality was checked using Shapiro-Wilk's test.The normally distributed parameters were analyzed using analysis of variance (ANOVA).A two-way ANOVA was used to test the significance of the treatments on the normally distributed parameters of vegetation and soil, while the mean treatment effects were analyzed using Tukey's HSD test.The non-parametric Kruskal-Wallis and Mann-Whitney tests were used for the data that were not normally distributed.All data were processed with Statistica software.

Meteorological variables
The experimental site was characterized by a sub-humid climate, with a long-term  mean annual rainfall of about 824 mm and average minimum and maximum temperatures of −8.0°C and 34.2°C, respectively.Over the 3 years of the experiment, air temperature ranged from −5.6 to + 36.8°C(Figure 1a), the highest amount of rainfall was recorded in 2014 (1,311 mm), and the lowest one in 2015 (533 mm).Intermediate rainfall was recorded in 2016 (999 mm) (Figure 1b).The maximum solar radiation and ET0 were measured in June and July, whereas the lowest values in December (Figure 1b,c).

AMF root colonization (%)
The root systems of all crops were colonized by AMF, increasingly from the first to the third year, except Jerusalem artichoke that showed an opposite trend (Table 4).AMF colonization also increased in the non-inoculated plants.Sorghum and miscanthus did not show root colonization in the first growing season.On average for all species, the roots of the inoculated plants were largely colonized in the first 2 years as indicated by the high percentages of the mean number of colonized fragments (F%), of the intensity of mycorrhization (M%) and of the abundance of arbuscules (A%), but no difference was observed in the third year (Table 4).

Biomass yield
AMF inoculation did not significantly influence the biomass yield of any crop.Despite AMF inoculation, the highest and lowest yearly average aboveground dry biomass yields were obtained with giant reed (42.7 ± 3.73 Mg ha −1 ) and perennial ryegrass (10.9 ± 0.39 Mg ha −1 ).Different biomass production trends were observed depending on the species over time (Figure 2).A significant yearly aboveground dry biomass increase was shown for giant reed and perennial ryegrass.The greatest miscanthus yield was obtained in the third cropping season, without any difference between the first and second seasons.Jerusalem artichoke followed an opposite trend: productivity declined from the first to the second year and then stabilized.The two annual crops maize and sorghum had the highest productivity in the first year.

Nitrogen and phosphorus biomass concentrations and uptake
AMF inoculation did not have any effect on the N and P biomass concentrations of the crops, but significant differences (Kruskal-Wallis test, p < 0.05) were observed among crops: perennial ryegrass biomass was characterized by the highest median concentration of both N (1.98%) and P (0.30%), followed by sorghum (Figure 3).The lowest biomass concentrations were found for miscanthus, with median values of 0.66% and 0.07% for N and P, respectively.AMF inoculation did not have any effect either on N uptake by any of the crops over time.The crops showed different N uptake levels (Figure 4).The highest levels occurred in the first growing season for Jerusalem artichoke, maize and sorghum, and then no difference was observed between 2015 and 2016; miscanthus followed an opposite trend, with the highest uptake in the last growing season.N uptake by giant reed increased exponentially from the first to the third cropping season.Perennial ryegrass was the only species that exhibited quite a stable N uptake throughout the experiment.At the end of the 3 years, the highest value of cumulative N uptake was calculated for giant reed (Figure 4), which removed from the soil about two times (1,548 kg N ha −1 ) the N supplied with DLF.Higher N uptake than supplied N was also found for sorghum (+140%) and Jerusalem artichoke (+123%).Maize was the only species able to take up almost all the N supplied by the DLF (98%).Miscanthus and perennial ryegrass took up the lowest amounts of N compared to the other crops.
Considering the yearly P uptake, again no effect was promoted by AMF, and the perennial crops showed the highest values in the third year, except Jerusalem artichoke that did not show any difference between the first and third years (Figure 5).An opposite trend was observed for the annual crops, which had the maximum uptake in the first year.These crops also showed the maximum cumulative P uptake (149.5 kg on average, about 1.7 times higher than the TP supplied with the DLF).Among the perennial species, the cumulative P uptake by giant reed was about 33% higher than the supplied dose, those of Jerusalem artichoke and perennial ryegrass were equal to the TP supplied with the DLF, while miscanthus showed a cumulative P uptake about 31% lower than the quantity supplied with the DLF (Figure 5).

Soil moisture and physico-chemical changes
Table 5 reports the average soil moisture values monitored in the first 2 years.For each crop, the moisture level in the first 0-90 cm of soil followed the meteorological trend, with a significantly lower value in 2015-1016 (lowest mean rainfall) than in 2014-2015 (highest mean rainfall).AMF inoculation did not influence the soil moisture in the boxes cultivated with maize.For the other crops, AMF inoculation significantly reduced the soil moisture in the first 0-30 cm layer for giant reed, while it increased in miscanthus and sorghum.In the 60-90 cm layer, AMF decreased soil moisture for Jerusalem artichoke and increased it for perennial ryegrass.
At the end of the experiment, AMF inoculation had significantly (p < 0.001) decreased soil Ca 2+ (−2.5%) and increased Mg 2+ , TP and TN (+1.8%, +4.8% and +8.6%, respectively) compared to the non-inoculated control soil (73.0 g Ca 2+ kg −1 , 33.5 g Mg 2+ kg −1 , 1.1 g TP kg −1 , and 0.18% TN).AMF inoculation had no significant effect on the soil organic carbon (SOC).This last variable was significantly affected (p < 0.01) by the crop.The highest value (1.59%) was reached after perennial  ryegrass cultivation, and the lowest one (1.40%) after giant reed and miscanthus cultivation.The SOC in the presence of the other three species was intermediate (1.46%).
Despite the treatments, the mean soil Mg 2+ concentration had decreased by 5.1% after 3 years, whereas the mean Ca 2+ , K + and Na + concentrations had increased by 2.5%, 13.5% and 76.9%, respectively, compared to the baseline soil concentrations (Table 1).

Nitrogen and phosphorus concentrations in percolation water
AMF inoculation significantly (Mann-Whitney test, p < 0.01) reduced the NH 4 -N (−32.8%) and increased the NO 3 -N (+70.0%)concentrations in the percolation water for all crops taken together (Figure 6).No specific effect of the crop species was observed on the NH 4 -N concentration in the percolation water over the years, with a median value of 1.39 mg NH 4 -N L −1 (Figure 6a) but the crop species had a significant (p < 0.001) effect on the NO 3 -N concentration (Figure 6b): the highest median value was found with maize (30.0 mg NO 3 -N L −1 ) and the lowest one with perennial ryegrass cultivation (1.21 mg NO 3 -N L −1 ).No difference was found between giant reed and sorghum (median value 17.9 mg NO 3 -N L −1 ), or between miscanthus and Jerusalem artichoke (median value 21.9 mg NO 3 -N L −1 ).TN was not affected by the studied factors, with a median concentration of 3.74 mg TN L −1 .
The P concentration in percolation water ranged from 0.0 to 1.83 mg L −1 for TP and from 0.0 to 1.35 mg L −1 for PO 4 -P.

Discussion
Several studies on giant reed and miscanthus have reported variable aboveground dry biomass and nutrient concentrations in tissues mainly depending on crop age and harvest time (Nassi o Di Nasso  The abundant rainfalls during the first year promoted Jerusalem artichoke growth as this species is sensitive to water stress, with a negative effect on tuber yield and plant biomass production (Schittenhelm 1999;Monti et al. 2005).Our results agreed with these studies, confirming the aboveground dry biomass yield reduction in the dry cropping seasons.The dry biomass yield (21.6 Mg ha −1 ) was in line with Curt et al. (2006)    findings of Gao et al. (2020), while slightly higher values were found in Sawicka et al. (2015) for plants fertilized with 200 kg N ha −1 (N: 2.7%, P: 0.28%).
The positive effect on abundant rainfall in the first year was also found for both maize and sorghum.In the following 2 years, the total aboveground biomass production was drastically reduced in both crops, particularly during the driest second year.When comparing the annual dry biomass yield per year, maize and sorghum showed similar production rates in 2014, but sorghum showed a higher yield (+14.0% in 2015 and + 40.0% in 2016) than maize in the following 2 years.Indeed, while maize is sensitive to water stress (Song et al. 2019), sorghum is a drought-tolerant crop (Martínez-Goñi et al. 2023).The average aboveground dry biomass yield of maize over the 3 years of the study (22.5 Mg ha −1 ) agreed with data reported by Maucieri et al. (2016) (22.0 Mg ha −1 ) and Ra et al. (2012) (20.1 Mg ha −1 ).The aboveground dry biomass yield of sorghum (26.2 Mg ha −1 ) was instead higher than in previous studies (Farré and Faci 2006, 18.3 Mg ha −1 ; Ra et al. 2012, 25.3 Mg ha −1 ).
Perennial ryegrass was cut five times throughout the experiment.The lowest total aboveground dry matter observed in the first two cuts in 2014 can be explained by the crop establishment period.The highest total aboveground dry matter was found after this period, at the third (2015) and fourth (2016) cuts.Perennial ryegrass is negatively affected by drought stress (Liu and Jiang 2010) and requires a large amount of water to sustain its growth (Turner et al. 2012).For this reason, we did not schedule another cut in October 2015 after the third cut because the drought conditions in the second year reduced and put a stop to shoot growth (Rogers et al. 2022).
The positive effect of AMF on plant growth has been widely reported, but with effects variable and dependent on the plant-mycorrhizal fungi interaction.Although several studies have reported AMF effectiveness in increasing biomass yields of the species studied in this research (Celebi et al. 2010;Firmin et al. 2015;Tauler and Baraza 2015;Selvakumar et al. 2016;Romero-Munar et al. 2017), in our case AMF inoculation did not support crops' yield increase.The main explanations could be: i) the low root colonization in the first year because of the physical disturbance of the soil when each box was filled with new soil (Van der Heyde et al. 2017); ii) the dry summer season conditions occurred during the second year (Silva et al. 2015); iii) the high nutrients input supplied with DLF (especially N) that reduced carbon allocation from the plants to the mycorrhizas (Berruti et al. 2014); iv) colonization from autochthones AMF which hide the effect only due to AMF inoculation.
Considering the environmental side of the present experiment focusing on the effectiveness of AMF on preventing N losses, AMF inoculation resulted in a higher NO 3 -N and lower NH 4 -N concentration in the percolation water.This result could be explained by the ability of AMF to rapidly deliver NH 4 -N to the plants, while apparently lacking the capacity to transfer NO 3 -N (Tanaka and Yano 2005).Our results are not in agreement with those reported by Bender et al. (2015), who did not find any effect of AMF on NH 4 -N leaching losses and N plant uptake when the soil NH 4 -N content increased.Comparing the crops tested in our experiment, the highest and lowest NO 3 -N concentrations in the percolation water were observed in the presence of maize and perennial ryegrass, respectively.Although these crops belong to the same botanical family (Poaceae), they exhibit different spatial, density, length, hair and relative exploration of the topsoil volume by the roots.The lowest N concentration found in the percolated water of perennial ryegrass can be explained by its perennial nature.
AMF have been reported to increase the SOC (Fall et al. 2022).Inoculation did not significantly change the SOC concentration in our experiment, probably because it did not significantly influence the aboveground dry biomass yield and in turn the belowground dry biomass.The higher SOC under perennial ryegrass cultivation than under giant reed and miscanthus cultivation can be attributed to the different root systems.Ryegrass is characterized by fine roots, whereas the other two plants produce rhizomes that may not yet have reached the turnover stage in 3 years.The higher soil monovalent cation concentration (especially Na + , +76.9%) than at the beginning of the experiment suggests that special attention should be paid to this parameter, especially in heavy soils fertilized with DLF for a long period.
The adoption of a circular economy approach in the bioenergy chain is relevant in a climate change scenario (Yang et al. 2022).This paper proposes the valorization of DLF in a close loop, as organic fertilizer of different bioenergy crops managed as for CH 4 production through anaerobic digestion.Our data indicated that the estimated CH 4 yield per hectare was significantly affected by the crop and the yield per year, even though the data are presented as averaged over the 3 years to make comparisons easier and the crops were harvested at different times and frequencies.Giant reed was the best biomass for biogas production per unit of surface, in spite of its lower intrinsic biomethanation potential (Marchetti et al. 2016).The expected CH 4 yield reported in this work is slightly higher than those presented by Dragoni et al. (2015) (9,900 Nm 3 ha −1 ) and Ragaglini et al. (2014) (9,580 Nm 3 ha −1 ).The other crops presented lower yields, as already confirmed by literature data, fitting the reference range values taken for maize (4,500-9,000 Nm 3 ha −1 ), which is considered as a reference crop for biogas production (Mayer et al. 2014;Vitez et al. 2021).It is worth noting that yields of up to 6,000 Nm 3 ha −1 have been reported for miscanthus by Mayer et al. (2014) and Kiesel and Lewandowski (2017).In our case, the higher CH 4 values could be linked to the dry aboveground biomass yield more than to the quality of the biomass.Sorghum was harvested several times.This fact makes a direct comparison with literature data hard, especially considering that the biomass yield is a consistent driver of the CH 4 yield (Dragoni et al. 2017).Our data were consistently higher than those reported by Kerckhoffs et al. (2014) whose experiment was based on a similar cultivar (Sugargraze) to the one used in the present work (Sugargraze II) but managed with a single harvest (~8,700 vs. 6,559 Nm 3 ha −1 ).The CH 4 yield of Jerusalem artichocke estimated in the present work (5,475 Nm 3 ha −1 ) based on an average yearly aboveground dry biomass yield of ~19.7 Mg ha −1 is slightly higher than the range reported by Lehtomäki et al. (2008) (3,100-5,400 Nm 3 CH 4 ha −1 ; aboveground dry biomass 9-16 Mg ha −1 ).The estimated CH 4 yield of perennial ryegrass (3,090 Nm 3 ha −1 ) was similar to those found by Sieling et al. (2013) (4,251 Nm 3 ha −1 ) under comparable conditions.

Conclusions
In a perspective of a closed-loop nutrients valorization in agriculture, our experiment focused on bioenergy crops, spotlighted giant reed as the most promising although careful management of nutrients balance is suggested on medium long term in order to avoid soil nutrients depletion.Moreover, if the use of DLF as an organic fertilizer for biomass production can be considered a viable alternative to reduce mineral fertilization inputs, a strict monitoring action should be focused on soil Na + concentration in the long-term period, especially if rainfed conditions do not allow a sufficient leaching.All the tested crops produced dry aboveground biomass in agreement with literature values regardless of the contribution of AMF inoculation.
The experiment was carried out from January 2014 to November 2016 at the L. Toniolo experimental farm of the University of Padova at Legnaro (45° 21' N; 11° 58' E; 6 m a.s.l.), north-east Italy.It was set up in order to compare six different crops -annual and perennial ones -under a single level of fertilization supplied by the DLF in the presence or absence of AMF inoculation (AMF-Y: presence of AMF inoculation; AMF-N: absence of AMF inoculation).The perennial crops were Arundo donax L. (local ecotype, giant reed), Miscanthus × giganteus Greef et Deu (local ecotype, miscanthus), Heliantus tuberosus L. (local ecotype, Jerusalem artichoke), Lolium perenne L. (cv.Mathilde, perennial ryegrass), while the annual crops were Zea mays L. (DKC 5401, maize) and Sorghum bicolor (L.) Moench × sudanense Stapf.(cv.Sugar Graze II, sorghum).

Figure 1 .
Figure 1.Meteorological data recorded during the trial: a) Maximum, minimum and average temperature; b) Rainfall and solar radiation; c) Potential evapotranspiration (ET 0 ).

Figure 2 .
Figure 2. Yearly dry aboveground biomass yields of the crops.Different lowercase letters indicate significant differences across years for each species (p < 0.05; Tukey's HSD test).Different uppercase red letters indicate significantly different three-year cumulative biomass yields among species (p < 0.05; Tukey's HSD test).

Figure 4 .
Figure 4. Nitrogen uptake by the crops.Different lowercase red letters indicate significant differences across years for each species (p < 0.05; Tukey's HSD test).Different uppercase letters indicate significantly different three-year cumulative N uptake by the species (p < 0.05; Tukey's HSD test).The horizontal red line represents the quantity of N supplied by DLF (digestate liquid fraction) in the three years.

Figure 5 .
Figure 5. Phosphorus uptake by the crops.Different lowercase red letters indicate significant differences across years for each species (p < 0.05; Tukey's HSD test).Different uppercase letters indicate significantly different three-year cumulative P uptake by the species (p < 0.05; Tukey's HSD test).The horizontal red line represents the quantity of P supplied by DLF (digestate liquid fraction) in the three years.
et al. 2011).The aboveground dry biomass yield found in this work for giant reed and miscanthus in the first 3 years of cultivation and harvested at the beginning of October are similar to those found byRoncucci et al. (2015),Giannini et al. (2017) andMaucieri et al. (2019).Regarding the nutrient contents, the giant reed N values found in the present work are quite high (>1%) and comparable to those found byBorin et al. (2013),Giannini et al. (2017) andMaucieri et al. (2018), where these species were cropped under unlimited nutrient availability in the first year.Both N and P miscanthus concentrations found in this study are similar to those reported byRoncucci et al. (2015) andGiannini  al. (2017)  for comparable harvesting dates.
(21.8 Mg ha −1 ) and Matías et al. (2013) (22.7 Mg ha −1 ).The aboveground biomass N concentration found in the present work was in line with the

Figure 6 .
Figure 6.Effect of AMF inoculation on a) the NH 4 -N concentration and b) the NO 3 -N concentration in the percolated water.Different letters show statistical differences at p < 0.01 (Mann-Whitney's test).

Figure 7 .
Figure 7. Methane yields per hectare obtained by averaging the aboveground dry biomass yield over three years by the reference value of methane potential per crop.Different letters indicate significantly different methane yields among species (p < 0.05; Tukey's HSD test).

Table 1 .
Soil chemical-physical characteristics at the experimental beginning.

Table 2 .
Chemical-physical characteristics of the digestate liquid fraction (DLF) applied during the three experimental years.

Table 3 .
Reference values used for the estimation of methane yield in the present study (*averaged value between the two cited in the article and deriving from two different measurement methods).

Table 5 .
Soil moisture (%) in the first two experimental years.Different letters indicate significant differences between years or between arbuscular mycorrhizal fungi (AMF) inoculation at p < 0.01 by Tukey's HSD test.