Effects of Faba Bean Strip Cropping in an Outdoor Organic Tomato System on Soil Nutrient Availability, Production, and N Budget under Different Fertilizations

Crop diversification has been identified as a tool to improve both productive and environmental performances of organic horticulture. We tested the introduction of faba beans in a tomato cropping system—both as preceding crop and in strip cropping (SC)—under different fertilization strategies—faba residues, animal manure, and vegetable compost. We studied: (i) the tomato nutrient uptake and yield and quality; (ii) the soil-N and P, the N-budget, and the mycorrhizal colonization. SC did not provide consistent positive effects on tomato production and quality, namely the N-uptake, N in the tomato yield, the mean fruit weight, and the sugar accumulation in berries, regardless of the type of fertilizers applied. SC improved the tomato dry yield and P-uptake, especially in years when the faba growth and the subsequent yield were problematic. Faba residues could provide sufficient N to sustain tomato production but care should be given in balancing additional N-inputs. Organic fertilization increased the soil-N concentration but did not always translate into an increased yield and a higher quality production, with the risk of escalating N-losses. SC improved the soil-P availability and mycorrhizal colonization in tomato, due to the rhizobia–mycorrhiza-mycorrhiza association, especially when coupled with organic fertilization. Finally, introducing faba as SC holds potential to improve the productive and environmental performance of organic tomato production.


Introduction
Combining crop production with positive environmental externalities represents the major challenge for the modern agricultural sector [1]. The demand for agricultural products from sustainable farming has risen consistently worldwide and the organic agricultural market has increased steadily in the last decades [2][3][4]. At the field level it has been shown that the introduction of organic management can provide a wide set of Ecosystem Services (ES) especially when agro-ecological practices are in place [5]. The need for organic management practices able to leverage ecological processes rather than focusing on substituting agrochemicals with organic inputs has strongly emerged in the debate concerning organic agriculture 3.0 [6,7]. Among those practices, crop diversification in time (e.g., crop rotation) and space (e.g., multicropping), has been shown to be able to stimulate positive ecological processes with variable effects on crop production and ES [8]. Diversification strategies have been also identified as a valuable tool to increase the economic and climatic resilience of farming systems [9,10]. Still, the variable effects of crop diversification on ecological processes call for agronomic and environmental assessments in order to adapt those strategies to specific pedo-climatic, environmental contexts [11].
Tomato represents a key crop of the European agricultural sector, with more than 16 million tonnes and Italy, Spain, and Portugal representing about 70% of the EU production [12,13]. Organic tomato production has rapidly developed in the last decades and is expected to grow further in the following years. Although crop rotation is mandatory according to the organic certification standards, other crop diversification strategies such as intercropping or strip cropping remain neglected in European organic tomato production systems. Particularly, introducing legumes in the crop rotation and/or as companion crops can bring several advantages in terms of, but are not limited to, soil nutrient availability, thereby offering the opportunity to decrease external inputs and mitigate possible environmental externalities [14]. Legumes can fix atmospheric-N and transfer N to the tomato crop, thereby resulting in a net N input to the system [15]. Moreover, the role of legumes in solubilizing unavailable soil P has been widely documented, especially in cereal-legume intercropping [16,17]. P represents an essential macronutrient for crops and is at the same time a non-renewable resource [18]. Unlocking unavailable soil P can therefore be considered a critical ecological function of legumes towards the sustainability of the whole cropping system: in Mediterranean systems, where high soil pH often determines a strong P insolubilization, legumes can play a role due to their high-exchange-capacity of root systems [19]. In addition, it has been shown that intercropping legumes with mycorrhiza host plants can improve P uptake and N transfer due to the synergistic association between the mycorrhiza and rhizobium [20,21]. However, very few papers have investigated the introduction of legumes as preceding or companion crops of organic tomato. Gatsios et al. (2021) [22] studied the effect of faba bean intercropping and cowpea as preceding crop to test whether these practices could replace farmyard manure or compost in terms of N input in organic greenhouse. Although green manure contributed substantial N, the authors concluded that legumes could not replace farmyard manure in the provision of N, mainly due to the low aboveground biomass produced by the legume. Different results were obtained by Salgado et al. (2021) [23], who tested the intercropping of different legumes in an organic cherry tomato system. The authors confirmed the capacity of legumes to fix atmospheric N and to transfer this N to cherry tomatoes with positive effects on growth and development.
In this study we focused on crop diversification in a Mediterranean organic tomato production system in relation to N and P availabilities along with productive parameters. Specifically, we compared faba bean-tomato crop rotation with strip cropping, including two fertilization strategies (stabilized farmyard manure and green compost) with the aim of providing an (i) agronomic assessment, by studying tomato N and P uptake, production, and quality; and (ii) a soil nutrient and environmental assessment by monitoring TOC, inorganic-N, nitrate-N, N budget, soil available P, and tomato root mycorrhizal colonization.

Study Site and Experimental Design
The experiment was conducted in 2018 and 2019 in the Monsampolo Vegetables organic certified Long-Term Experiment (MOVE.LTE) located in Monsampolo del Tronto (AP), Italy (42 • 53 N, 13 • 48 E). Climate is classified as thermo-Mediterranean [24] with annual precipitations accounting for about 564 mm and average air temperature for 9 • C and 20 • C in October-March and April-September, respectively. Soil is classified as typic Calcixerepts fine-loamy and mixed thermic [25]. The three-blocks experimental design compared monocropped (MC) faba bean (Vicia faba L., local cv "Fratterosa", FB) and tomato (Solanum Lycopersicum L., local cv "SAAB CRA", T), to faba bean-tomato strip-cropping (SC) in 2.8 m × 3.7 m plots ( Figure 1). Row distance accounted for 0.7 m for both crops, while plant distance accounted for 0.2 and 0.5 for faba and tomatoes, respectively. Faba was sown in November after ploughing and arrowing in all the experimental plots. Following the harvest of the fresh faba pods (first week of May 2018 and 2019), faba bean was terminated by flattening using a roller crimper in the plots where the tomato would have been transplanted (tomato MC plots and tomato SC strips). In all the faba bean MC plots and in SC strips, the legume remained in field up to half July 2018 and 2019 until dry grain harvesting, while in tomato MC plots and SC strips, the flattened faba bean residues (FF) formed a green mulch, where tomato was transplanted in rows. Consequently, N input from FF was always considered as additional N supply to N added with organic fertilizers to tomato plants.  Tomato plants were drip irrigated in summer (300 L m −2 ). Three fertilizer treatments were applied on tomato MC and SC plots: (i) organic fertilizer based on animal manure at 120 kg N ha −1 applied in two times, mid-May and mid-June (FYM+FF); (ii) vegetable compost at 120 kg N ha −1 applied at the end of May (VEGco+FF); and (iii) no fertilization (control), where N supply was derived only by flattened faba bean residues (FF). Faba bean was not fertilized. The chemical characteristics of the FYM+FF and vegetable composts are reported in Table 1. The experimental field was shifted every year to comply with the organic certification standards of the MOVELTE site.  Tomato plants were drip irrigated in summer (300 L m −2 ). Three fertilizer treatments were applied on tomato MC and SC plots: (i) organic fertilizer based on animal manure at 120 kg N ha −1 applied in two times, mid-May and mid-June (FYM+FF); (ii) vegetable compost at 120 kg N ha −1 applied at the end of May (VEGco+FF); and (iii) no fertilization (control), where N supply was derived only by flattened faba bean residues (FF). Faba bean was not fertilized. The chemical characteristics of the FYM+FF and vegetable composts are reported in Table 1. The experimental field was shifted every year to comply with the organic certification standards of the MOVELTE site.

Plant Analysis
Total tomato yield was measured by adding the marketable production and the waste production (undersized fruits, fruits attacked by pathogens, and fruits with physiological alterations). The fruits were harvested at commercial ripeness in the first six trusses of the plant in July and August. After harvesting, fruits were counted, weighed, and the percentage of soluble solid residue ( • Brix) was determined with a PAL-1 pocket digital refractometer (ATAGO Co., Ltd., Tokyo, Japan).
Plant P content on dry matter was determined on 2 g of tomato fresh biomass, incinerated at 400 • C for 24 h and re-solubilized by HCl 0.1 M [26] using simultaneous plasma emission spectrophotometer (ICP-OES Iris; Thermo Optek, Milano, Italy).
The faba bean and tomato roots were sampled at field by using stainless steel cylinders of 6 cm diameter and 20 cm length [27]. Three root subsamples per plot were collected, then pooled to obtain a tomato root sample per treatments per block. Roots were discarded from the rhizosphere soil by washing in distilled water in a sieve of 0.5 mm mesh, and then ordinated into first, second, and third-order lateral roots for further analyses, using the staining method described by Grace and Stribely (1991) [28] and Trinchera et al. (2021) [27]. Root fragments were placed on grinded slides, mounted in a drop of glycerol, and observed under a light microscope (Nikon E100) at 40X magnification.
Tomato root fragments were placed on grinded slides, mounted in a drop of glycerol, and observed under a light microscope (Nikon E100). The mycorrhizal colonization intensity (M%) of tomato and faba bean roots was assessed by applying the method of Trouvelot et al. (1986) [29], based on the observation on of the root fragments occupied by AMF structures, thus calculating the mycorrhizal colonization intensity (M%) [30].
To understand the P dynamic in soil as affected by rhizobia and mycorrhizal fungi, on 4 June 2019 the faba bean roots at 0-30 cm were sampled in MC and SC plots (only in correspondence with tomato SC fertilized with VEGco+FF fertilization). The same procedure was followed, collecting first, second, and third-order lateral roots to study the surface of root nodules. After a preliminary observation under a light microscope, the root nodules were observed under Scanning Electron Microscope (EVO MA10 Zeiss) at variable pressure (selected at 10 Pa), equipped with a LaB 6 lamp to increase image brightness. Secondary electrons detector (SE) was used to better study the morphology of nodule surface and verify the presence of mycorrhizal hyphae on nodules' surface, with the contemporary presence of rhizobia colonization.

Soil Analysis
At tomato harvest (July 2018 and 2019), three replicates per block of bulk soil were sampled using an auger at 0-30 cm depth in both monocropping and strip-cropping tomato plots under FF, VEGco+FF, and FYM+FF. Samples were, air dried, homogenized by sieving (2 mm mesh size) to remove fine roots and large organic debris and stored at room temperature until use. For soil inorganic-N, a subsample each was taken after mixing each sample into plastic bags in order to homogenize and stored at −20 • until the extraction.
Soil inorganic-N and nitrate-N were sampled three times per year at planting, midseason, and harvest, with three replicates per block. Soil total inorganic N (N-NO 3 + N-NH 4 ) was determined after soil extraction by 2 M KCl (1:10 w/v) and measured by continual flow colorimetry according to Krom (1980) [31] and Henriksen and Selmer-Olsen (1970) [32] for N-NH 4+ and N-NO 3 , respectively.
Soil organic C was determined on 0.250 g of dried soil using LECO TOC Analyzer (mod. RC-612; LECO Corporation, St. Joseph, MI, USA), while total N content was assessed by using LECO Nitrogen analyser FP-528, LECO Corporation, St. Joseph, MI, USA.
Soil available P, on dry soil, was determined on 5 g of dry soil using Mehlich III method [33], then analyzing the obtained acidic solution by using simultaneous plasma emission spectrophotometer (ICP-OES Iris; Thermo Optek, Milano, Italy).

Nitrogen Budget
N-budget was calculated using a mass balance approach using Equation (1).
where: Fr: N in faba bean residues; Fert: N applied through FYM+FF and green compost; Ntr: N in tomato residues; Nty: N in tomato yield; ∆Sn: difference between total soil N at planting and harvesting.

Statistical Analysis
Mixed effect models were used to test the effect of treatments (fertilization and cropping systems) on plant and soil parameters. We tested four models to include all the possible interactions across the factors: CS, fertilization, and year and we selected the models with the lower BIC. Afterwards, we included the SOC at planting to account for the within and between plots soil variability. We kept SOC at planting in the model when it further decreased the BIC of the selected model. Furthermore, year was nested into block in the random part of the model as means to consider any difference in experimental fields between years (lme4 package). The percentage of mycorrhizal colonization was analyzed through beta-regression model (betareg). In all cases, residuals were visually assessed and a Shapiro-Wilk test was performed. Analysis of variance (type II SS and type III in case of significant interactions) was used to check for statistically significant factors. Estimated marginal means (emmeans package) were used to obtain p-value corrections, with Tukey's post-hoc test (α = 0.05). All statistical analyses were performed in R (version 4.1.0, 2021). Figure 2 shows the climatic data of the two experimental years along with the historical data averaged over 30 years. We observed a large difference in precipitation between 2017-2018 and 2018-2019. Overall rainfall was more than 1.5 times higher in 2017-2018 (total precipitation = 1020 mm) compared to 2018-2019 (total precipitation = 670 mm). Moreover, a snowfall and a severe frost at the beginning of February 2018 further characterized a very wet year, with a consequent negative effect on the faba bean growth and production. Indeed, anthracnose (Ascochyta fabae Speg.) and chocolate spot (Botrytis fabae) took advantage of the high temperature, causing serious damage to the faba bean. 3.2. Agronomic Assessment: N and P Uptake, Tomato Production, and Quality Table 2 shows the results of the analysis of variance for the tomato production parameters, as affected by cropping system (CS), fertilization strategy (Fert) and year. No effects of the factors considered in the analysis were found for N in yield, N uptake, P uptake, fresh yield, and sugar content (°Brix). Cropping system (p ≤ 0.01) and the CS ×  Table 2 shows the results of the analysis of variance for the tomato production parameters, as affected by cropping system (CS), fertilization strategy (Fert) and year. No effects of the factors considered in the analysis were found for N in yield, N uptake, P uptake, fresh yield, and sugar content ( • Brix). Cropping system (p ≤ 0.01) and the CS × year interaction (p ≤ 0.05) significantly influenced P recovery in the yield. We found a higher P in the tomato total yield under SC (mean available-P = 10.4 kg ha −1 ) compared to MC in 2018 (mean available-P = 4.6 kg ha −1 ), while no significant differences were found between the two cropping systems in 2019 (Figure 3). Similarly, the tomato dry yield was significantly higher in 2018 under strip cropping (mean tomato dry yield = 2.8 kg of DM ha −1 ), compared to MC (mean tomato dry yield = 1.6 kg of DM ha −1 ) in 2018, while no differences were found between the two cropping systems in 2019 ( Figure 4). The mean fruit weight only differed across years (p ≤ 0.05). We did not find any significant effect of the cropping system nor of the fertilization strategy on the tomato fresh yield, N in yield, N uptake, and sugar accumulation.      Table 3 shows the results of the analysis of variance for TOC, soil inorganic trate-N, available-P, and mycorrhizal colonization intensity (M%). As expected, T not change significantly across the experimental years and the cropping systems 3). Concerning N, we found a significant triple interaction between CS × Time × y 0.001) with the highest total inorganic-N and nitrate-N under strip cropping in mid only in 2018 (data not shown). We also found a significant effect of fertilization str on both total inorganic-N and nitrate-N (Figure 5a,b). As expected, the applica FYM+FF (mean total inorganic-N = 12.7 mg kg −1 ) and VEGco+FF (mean total inorg = 12.5 mg kg −1 ) significantly increased the inorganic-N compared with the FF (mea ganic-N = 10.3 mg kg −1 ). The soil Total inorganic-N increased, irrespectively of the of the N input, as the two treatments did not differ significantly. A different scena observed for nitrate-N, where the FYM+FF application significantly increased the N (mean nitrate-N = 5.0 mg kg −1 ) compared to the FF (mean nitrate-N = 3.6 mg k not in comparison with VEGco+FF (mean nitrate-N= 4.7 mg kg −1 ) (Figure 5b).   Table 3 shows the results of the analysis of variance for TOC, soil inorganic-N, nitrate-N, available-P, and mycorrhizal colonization intensity (M%). As expected, TOC did not change significantly across the experimental years and the cropping systems (Table 3). Concerning N, we found a significant triple interaction between CS × Time × year (p ≤ 0.001) with the highest total inorganic-N and nitrate-N under strip cropping in midseason only in 2018 (data not shown). We also found a significant effect of fertilization strategies on both total inorganic-N and nitrate-N (Figure 5a,b). As expected, the application of FYM+FF (mean total inorganic-N = 12.7 mg kg −1 ) and VEGco+FF (mean total inorganic-N = 12.5 mg kg −1 ) significantly increased the inorganic-N compared with the FF (mean inorganic-N = 10.3 mg kg −1 ). The soil Total inorganic-N increased, irrespectively of the quality of the N input, as the two treatments did not differ significantly. A different scenario was observed for nitrate-N, where the FYM+FF application significantly increased the nitrate-N (mean nitrate-N = 5.0 mg kg −1 ) compared to the FF (mean nitrate-N = 3.6 mg kg −1 ) but not in comparison with VEGco+FF (mean nitrate-N = 4.7 mg kg −1 ) (Figure 5b). Table 3. Results from the analysis of variance for tomato soil Total inorganic-N, nitrate-N, N balance and available-P, TOC, and Mycorrhizal colonization intensity. CS = Cropping system; Fert = fertilization.    (Figure 6a). In 2018, the high N input from the FF was more than sufficient to sustain tomato production: the further application of fertilizers increased the N-surplus to worrying levels, questioning the environmental sustainability of the cropping system. A different scenario was observed in 2019, when the N input from the FF was on average about one third lower than in 2018. In 2019, where fertilizers were not applied (FF, control plots), the N budgets were negative, indicating a depletion of the soil N pool. Still, in the case of fertilizers application, an N-surplus indicated an N accumulation that varied between 81-and 99 kg ha −1 . As expected, the application of organic fertilizers increased significantly the N budget, regardless of the type of fertilizers applied (Figure 6b, Table 4).  (Figure 6a). In 2018, the high N input from the FF was more than sufficient to sustain tomato production: the further application of fertilizers increased the N-surplus to worrying levels, questioning the environmental sustainability of the cropping system. A different scenario was observed in 2019, when the N input from the FF was on average about one third lower than in 2018. In 2019, where fertilizers were not applied (FF, control plots), the N budgets were negative, indicating a depletion of the soil N pool. Still, in the case of fertilizers application, an N-surplus indicated an N accumulation that varied between 81-and 99 kg ha −1 . As expected, the application of organic fertilizers increased significantly the N budget, regardless of the type of fertilizers applied (Figure 6b, Table 4).

Phosphorus, Mycorrhizal Colonization, and Faba Bean Root SEM Analysis
The available-P was significantly affected by the year (p ≤ 0.001) and the interaction CS x year (p ≤ 0.05) ( Table 3). We found a higher available-P under SC compared to MC in 2019, while no differences across cropping systems were found in 2018 (Figure 7).

Phosphorus, Mycorrhizal Colonization, and Faba Bean Root SEM Analysis
The available-P was significantly affected by the year (p ≤ 0.001) and the interaction CS × year (p ≤ 0.05) ( Table 3). We found a higher available-P under SC compared to MC in 2019, while no differences across cropping systems were found in 2018 (Figure 7). Mycorrhizal root colonization was significantly affected by CS (p ≤ 0.001), the fertilization strategy (p ≤ 0.01), and the year (p ≤ 0.001) ( Table 3). Mycorrhizal colonization more than doubled under SC (mean mycorrhizal colonization = 19.2%) compared to MC (mean mycorrhizal colonization = 8.3%) (Figure 8a). Similarly, we found a positive effect of organic fertilization on mycorrhizal colonization intensity with the fertilized plots (mean mycorrhizal colonization = 16.5%), showing twice as much root colonization as the non-fertilized fields (mean mycorrhizal colonization = 8.0%) (Figure 8b). Neverthless, the type of fertilizers did not significantly influence roots colonization by mycorrhiza. Finally, a significant higher mycorrhizal colonization was found in 2019 (mean mycorrhizal colonization = 20.2%) compared to 2018 (mean mycorrhizal colonization = 7.2%) (data not shown). Mycorrhizal root colonization was significantly affected by CS (p ≤ 0.001), the fertilization strategy (p ≤ 0.01), and the year (p ≤ 0.001) ( Table 3). Mycorrhizal colonization more than doubled under SC (mean mycorrhizal colonization = 19.2%) compared to MC (mean mycorrhizal colonization = 8.3%) (Figure 8a). Similarly, we found a positive effect of organic fertilization on mycorrhizal colonization intensity with the fertilized plots (mean mycorrhizal colonization = 16.5%), showing twice as much root colonization as the nonfertilized fields (mean mycorrhizal colonization = 8.0%) (Figure 8b). Neverthless, the type of fertilizers did not significantly influence roots colonization by mycorrhiza. Finally, a significant higher mycorrhizal colonization was found in 2019 (mean mycorrhizal colonization = 20.2%) compared to 2018 (mean mycorrhizal colonization = 7.2%) (data not shown). of organic fertilization on mycorrhizal colonization intensity with the fertilized plots (mean mycorrhizal colonization = 16.5%), showing twice as much root colonization as the non-fertilized fields (mean mycorrhizal colonization = 8.0%) (Figure 8b). Neverthless, the type of fertilizers did not significantly influence roots colonization by mycorrhiza. Finally, a significant higher mycorrhizal colonization was found in 2019 (mean mycorrhizal colonization = 20.2%) compared to 2018 (mean mycorrhizal colonization = 7.2%) (data not shown). Faba bean nodules colonized by rhizobia were observed by SEM to verify the presence of mycorrhizal hyphae on nodule surfaces. In Figure 9, the nodule surface of faba bean roots in monocropping and strip cropping were compared, focusing on endophytic rhizobia colonization and fungi hyphal development (fertilization: VEGco+FF). The mycorrhizal spore found close to external hyphae (spr), confirmed that mycorrhizal mycelial coverage took place on faba bean nodules (Figure 9d). Faba bean nodules colonized by rhizobia were observed by SEM to verify the presence of mycorrhizal hyphae on nodule surfaces. In Figure 9, the nodule surface of faba bean roots in monocropping and strip cropping were compared, focusing on endophytic rhizobia colonization and fungi hyphal development (fertilization: VEGco+FF). The mycorrhizal spore found close to external hyphae (spr), confirmed that mycorrhizal mycelial coverage took place on faba bean nodules (Figure 9d).

Agronomic Assessment: N and P Uptake, Tomato Production, and Quality
The introduction of legume crops has been often identified as a tool to recycle atmos-

Agronomic Assessment: N and P Uptake, Tomato Production, and Quality
The introduction of legume crops has been often identified as a tool to recycle atmospheric N into the soil, with positive effects on N uptake and crop production. Previous studies have estimated that N fixation represents about 50% of the total N in legume biomass and that part of this N can be transferred to the tomato [23]. Still, in our study we did not find any effect of CS on N uptake and N in tomato yield (Table 3). We provide three hypotheses to explain our results. Firstly, the faba bean under SC was going through senescence when the tomato was transplanted, and it is conceivable that the N fixation and the transfer to the tomato plant would have been at its minimum. Secondly, there might be a lack of synchrony between the availability of N from legume-flattened residues under degradation and the uptake of tomato. The synchrony between crop uptake patterns and N availability has been widely reported to be a critical issue in organic nutrient management [34,35]. As an example, Gatsius et al. (2021) [36] found an excessive mineral N concentration during the vegetative stage of tomato following the cultivation of different legumes as green manures. Afterwards, the mineral N dropped at concentrations lower than the optimal resulting in an N deficiency of tomato. Thirdly, both CSs included faba bean as preceding crop and the N input from the flattened faba bean was comparable (Table 4), thereby reducing the differences between the treatments in terms of the N uptake and the N in yield.
A different scenario was observed for P uptake, which was significantly higher in tomato biomass under SC in 2018 as compared to MC (Figure 3). Different studies highlighted the capacity of legumes to select microbial rhizospheric communities and/or to produce root exudates able to solubilize P [37,38]. In our experiment we hypothesised that P was more successfully mobilized by the faba bean in SC compared to MC, but it was not uptaken by the legume, which was barely productive in 2018. Therefore, the tomato benefited by solubilized P that was accumulated in the plant biomass. Furthermore, synergistic relationships between rhizobium and mycorrhizal fungi were documented [39] and such an interaction could have further increased the P uptake by tomato.
The effect of crop diversification on tomato yield has been reported to be dependent on climatic conditions along with the characteristics and development of the companion legume crops [23]. Faba bean has been shown to provide yield advantages compared with other legumes (e.g., common bean) [36]. Still, in our experiment the higher P uptake by tomato in 2018 corresponded to a higher dry matter yield under SC but no differences were found in terms of fresh yield. A water deficit can increase dry matter in tomato production [40] while intercropping can trigger competitive relationships for water uptake. In our experiment, irrigation supplied sufficient water to maintain a fresh yield but still the water uptake by faba bean under SC may have affected the tomato stem water potential, thereby increasing the accumulation of dry matter in tomato yield. Despite the higher dry matter found in tomato yield under SC in 2018, we did not find an effect of CS on sugar concentration. Sugar concentration has been shown to be deeply affected by intercropping in greenhouse tomato production especially when fresh yield differences were observed [41]. Nevertheless, our results are comparable with Liu et al. (2014) [42] who found no difference in fresh tomato yield despite the higher dry matter, and no difference in sugar concentration in a tomato-garlic intercropping. Total organic carbon has been often considered as a proxy for soil health due to its capacity of improving chemical, physical, and biological soil fertility [43]. Farming practices that do not deplete and possibly increase TOC therefore play an important role in improving soils and guarantee healthy soils for future generations. Introducing legumes, and particularly faba bean, has been shown to increase TOC [44]. Nevertheless, we did not find any effect of cropping system or fertilization strategy on TOC. Firstly, increasing TOC can take more than a year [45]: we rotated the fields every year to comply with the organic certification standard and hence we could not monitor the TOC accumulation throughout the experimental years in the same fields. Secondly, faba bean was grown in both CSs, making it difficult to highlight differences across MC and SC.

Nutrient and Environmental
For the same reason, we did find minor differences in total inorganic-N and nitrate-N across CS throughout the years. Nevertheless, other studies where faba bean was grown as a preceding crop before tomato transplanting found a higher soil N compared to other legumes (cowpea and common bean) [36].
We also found a strong effect of fertilization strategy on total inorganic-N, regardless of the type of fertilizers. We would have expected a higher total inorganic-N under FYM+FF compared to VEGco+FF due to the higher mineralization potential. Probably, the presence of faba bean provided enough N to mitigate the potential N-immobilization, thereby reducing the difference in total inorganic-N between FYM+FF and VEGco+FF. Indeed, in the work by Chaloub et al. (2013) where no additional N input was applied, a buildup in soil organic-N was found, due to the application of more stable soil organic matter in compost compared to manure [46]. Still, in this experiment, the compost and farmyard manure application rates were equalized based on C, while in our study we focused on applying the same N rate regardless of the type of input.
As far as nitrate-N is concerned, compost application differed neither from FF nor from FYM+FF. This suggests that the compost application favored ammonium-N rather than nitrate-N compared to FYM+FF. Similar results were obtained in a maize alfa-alfa rotation and in a red raspberry experiment, where the authors found a higher nitrate-N and nitrate-N leaching under farmyard and poultry manure compared with compost [47,48].
The Nitrogen budget has been considered a critical agro-environmental indicator by policy makers, scientists, and agronomists [49]. In our system we found a high variability in the N input from faba bean residues across experimental years. Berry et al. 2003 [50] argued that the contribution of grain legumes on the N budget can be minor as most of the N fixed is exported in the grain harvest. In our experiment, N accumulated into the faba bean residues in 2018, when faba bean production was extremely low: in this year, faba bean residues sustained the N budget. The further addition of organic amendments escalated the N-surplus with potential environmental damages. However, in a year characterized by a regular faba production, the sole faba bean residues were not enough to sustain the N budget with a potential risk of soil N depletion. In this case, the N input from organic fertilizers was definitely beneficial, even though a reduction in the N rate should still be considered to mitigate possible N losses.

Available P and Mycorrhizal Colonization
Phosphorus is an essential element for tomato growth and production. P is not a mobile element in soils and it has been estimated that only 15-30% of the P applied via fertilizers is available for plants [51]. As already mentioned, crop diversification can improve soil P availability especially when legumes are grown as a companion crop due to the ability of legumes to solubilize P [52]. Overall faba bean has been indicated as a suitable P-mobilising legume in crop rotations to improve P availability for the following crop [38]. In our experiment we found a higher available P under SC compared to MC in 2019, while no differences were found in 2018 (Figure 7). At the same time we found a higher P uptake by tomato under SC in 2018 and no significant differences in 2019 (Figure 3). The faba bean helped in solubilizing unavailable P, especially under SC where the faba was not terminated between the tomato rows. Still in 2018, faba bean production was very low due to snowfall and disease attacks, thereby resulting in a low P uptake by the faba. This facilitated the P uptake by the tomato under SC (Figure 7) so that we did not find differences in soil available P among CS. Conversely, in 2019, faba bean production was almost four times higher (data not shown), so that the solubilization of P and the mineralization of faba bean residues occurred at a later stage, when P uptake from tomato was not at its peak. Indeed it has been reported that the maximum uptake of P in tomato plants is concentrated in the budding stage of the crop [53]. As a result, in our experiment, the available P could have been solubilized too late compared to the tomato P demand, thereby accumulating in the soil under SC compared to MC. Once again the synchrony between crop uptake and soil nutrient availability defines the efficiency and the sustainability of the organic farming system calling for site-specific optimization of the fertilization strategies.
Among the benefits brought by SC, we found a strong effect on tomato mycorrhizal colonization (Figure 8a). Legumes are recognized as an important plant functional group that can form a tripartite symbiosis with nitrogen-fixing Rhizobium bacteria and phosphorusacquiring arbuscular mycorrhizal fungi [54]. It is noticeable that in 2018, rhizobia nodules of faba bean roots, particularly under SC, were partially covered by mycorrhizal hyphal mycelium, confirming the contemporary rhizobial endophytic colonization of the legume and the symbiotic mycelium development on its roots (Figure 9c,d). This synergistic effect played by rhizobia and mycorrhizal fungi in faba bean evidently increased soil P mobilization and promoted the following mycorrhization of tomato, thus increasing Puptake. Our results are in line with the literature concerning faba bean-cereal association. Ingrafia et al. (2019) [55] found increased mycorrhizal root colonization when faba bean and wheat were intercropped compared to pure stand. In addition, the beneficial effects of faba bean intercropping with cereals in terms of P and N availability driven by mycorrhizal and rhizobium associations have been already reported [56,57]. In relation to vegetable crops also, it was already observed that the introduction of mycorrhzal preceding crops, such as winter-cereal cover crops, can increase the mycorrhization of melon via sharing a common mycelial network at the belowground, as happened in our faba bean tomato system [30].
We also found a strong effect of organic fertilization on mycorrhizal colonization on tomatoes, irrespectively of the type of fertilizers (Figure 8b). Bilalis et al. (2010) [58] have shown comparable results. They argued that the addition of compost and manure enhanced soil physical health that, in turn, was associated with a higher mycorrhizal colonization. An additional study presented the results from a long-term experiment that compared the effect of different fertilizers, namely mineral fertilizers, cattle manure, and manure with biodynamic preparations, on mycorrhiza colonization. The authors reported a higher mycorrhizal colonization under manure compared to plots amended with mineral fertilizers [59]: it was found that the input of manure and compost improved the habitat of mycorrhiza by improving physical, biological, and chemical soil health as compared to chemical fertilizers.

Conclusions
This study presents an agronomic and environmental assessment of the tomato-faba bean cropping system, where the legume was introduced both as preceding and companion crops under two different fertilization strategies.
From an agronomic point of view, SC did not provide consistent positive effects on tomato production and quality, namely N uptake, N in tomato yield, mean fruit weight and soluble solid residue in berries, regardless of the type of organic fertilizers applied. SC can anyway improve tomato dry yield and P uptake in years when faba bean growth and production are problematic.
As far as the nutrient management and environmental implications, our results suggest that care should be given in balancing N inputs in faba bean tomato cropping systems. N inputs from faba beans can provide a sufficient N to sustain tomato production but they can vary quiet remarkably, so that quick estimates are needed to balance additional fertilization. To this end, further studies to test and operationalize rapid and cost-effective monitoring systems of N input in multicropping systems are needed. Organic fertilization can also increase soil N concentration but does not always translate into a higher and more qualitative production, with the risk of escalating N losses. Although this study does not specifically address the synchrony between soil N availability and crop uptake, our results indicate that this is still a key topic to advance sustainable cropping systems in the organic sector. SC, also holds potential to improve soil P availability and mycorrhizal colonization in tomato cropping systems, with positive environmental and agronomic effects. The diversification of crops and organic fertilization stimulates the mycorrhizal fungi to increase the colonization of the roots, as a result of the sharing of the root systems between coexisting species in an optimal environment for the development of the hyphal mycelium, as well as of a nutritional status of the soil for fungal colonization. The rhizobia-mycorrhiza association suggested that the introduction of faba bean in organic SC tomato systems can potentially contribute to maintainin production while reducing external P inputs.
Finally, introducing faba beans in tomato production, especially as SC, holds the potential to improve the productive and environmental performance in Mediterranean-and also in other climatic conditions-systems but careful monitoring of nutrient flows is needed to fully exploit its potential.