**1. Introduction**

According to recent statistics, land managed under organic farming regulations in Europe has increased by almost 75% in the last decade [1]. Consumer demand for environmental sustainability as well as safety and food quality concerns continue to drive the organic industry and to encourage farmers to convert their agricultural systems to organic farming. However, organic producers rely primarily on intensive and frequent tillage for weed management, organic fertilizers and residue incorporation, and seedbed preparation [2], in a way that sometimes violates the objective of organic farming to sustain soil health. Intensive tillage reduces soil quality, facilitates erosion through the destruction of soil structure, increases loss of topsoil organic matter, and decreases soil biological activity and biodiversity [3]. No-tillage systems were developed a few decades ago in conventional agriculture to mitigate these problems and to provide economic savings by eliminating tillage and excessive traffic on fields [3–5]. Benefits to soil fertility and other ecological services (i.e., weed and pest suppression, nutrient cycling) are provided by cultivation of cover crops in rotation with cash crops as well [6]. Using legume species as cover crops also provides additional N fixed from the atmosphere into the agro-ecosystem thus improving N nutrition of the cash crop and increasing soil nitrogen organic pool [7].

Recently, researchers have been increasingly investigating cover-crop-based no-till (NT) as a sustainable practice to eliminate the reliance on mechanical tillage and maximize the benefits of cover crops and resource use e fficiency in organic farming [6,8]. In these systems, cover crops are terminated without incorporating residues into the soil, thus leaving a thick mulch into which the subsequent cash crop is planted. This requires the necessity to produce large cover crop biomass as well as a good managemen<sup>t</sup> of their residues to provide maximum weed suppression and nutrients adjustments, e.g., reduce immobilization, enhance N release and synchronization with plant needs [9]. Weed managemen<sup>t</sup> and nutrient availability are two factors known to challenge the performance of crops in organic no-till production. In such systems, weeds tend to increase with higher seedling recruitment in the upper soil layers and large infestations of perennial weeds [2,10]. Cover crops can reduce weed infestation during their growth and/or by their residues on soil surface making a physical barrier, preventing sunlight reaching the soil surface or through allelopathy [11]. With reduced or absence of tillage, mineralization of soil organic matter can also be slowed down which would make N a limiting factor in these conditions and compromise yield production [12–14].

Italy is the second largest producer of processing tomato after the USA with more than 72,000 ha dedicated to it as of 2018 [15,16]. In this study, we aimed to understand how the transition to no-till would impact the production of tomato and if a mulch of cover crop residues would be able to replace plastic mulch which is costly and di fficult to dispose of when the material is not biodegradable. To this end, the following field experiments (2015–2017) compared cover crop-based no-till and conventionally tilled systems for organic processing tomato production under Mediterranean conditions in terms of crop growth, yield, fruit quality, N uptake as well as the changes in soil nitrates, soil compaction, and weed infestation.

#### **2. Materials and Methods**

#### *2.1. Field and Treatments Description*

The experiments were conducted on certified organic fields at the Center for Agri- Environmental Research "Enrico Avanzi" of the University of Pisa (San Piero a Grado, Pisa, Italy) for two seasons (2015–2017). Seven systems were adopted: squarrose clover (*Trifolium squarrosum* L.) rolled, flamed, and followed by a direct transplantation of tomato (*Solanum lycopersicon* L. var. Elba F1, a processing cultivar that can be used also for fresh consumption) (NT-CC); squarrose clover rolled and flamed, followed by a direct transplantation of tomato and supplemented with weeding interventions, i.e., inter-row mowing (NT-CC-SW); squarrose clover incorporated as green manure (CT-CC); squarrose clover incorporated and the soil covered with black biodegradable plastic mulch (Mater-Bi ®) set over the season (CT-CC-PM); fallow conventionally-tilled soil covered with plastic mulch (CT-NC-PM), fallow conventionally-tilled with soil kept bare (CT-NC), and a weedy control left untilled with natural vegetation (NT-NC). The fields were moldboard ploughed and harrowed in 16 November 2015 and 3 October 2016 before the cover crop (*T. squarrosum*) broadcast manual sowing at 50 kg ha−<sup>1</sup> seeding rate on 17 December 2015 and 35 kg ha−<sup>1</sup> on 12 October 2016. The sowing densities of the clover di ffered across the two years because of the di fferent germination rate and of the delayed sowing date in 2015, but they were targeted to the same plant densities (667 plants m<sup>−</sup>2).

In conventionally tilled (CT) plots, the cover crop was terminated using a rotary hoe at around 15 cm depth. Fallow plots were prepared for transplanting the same way. In these plots, inter-row cultivation was also performed for subsequent weed control. The cover crop in NT treatments was terminated with two passes of a roller-crimper (Eco-roll, Clemens Technologies, Wittlich, Germany) followed by one pass of flaming (MAITO Srl., Arezzo, Italy based on a prototype designed and fully realized at the University of Pisa) to enhance cover crop devitalization [17] on 23 May 2016 and 10 May 2017. In NT plots with supplemental weeding, three inter-row mowing interventions (lawn mower) were done during the early season of tomato growth.

In plots without plastic mulch, tomato seedlings were transplanted on 23 June 2016 and 11 May 2017 at a density of 2.22 plants m<sup>−</sup><sup>2</sup> (0.3 m along the row, 1.5 m between the row) with a commercial vegetable transplanting machine ("Fast" model, Fedele costruzioni Meccaniche, Lanciano, Chieti, Italy) modified at the University of Pisa in order to be properly used both on tilled and untilled soil [18]. Tomato seedlings were instead manually transplanted at the same plant density on plastic mulch systems. The distance between tomato single rows (1.5 m), fine-tuned for the plastic mulch system, was kept the same for all treatments to avoid additional variability that can influence the results and conclusions. Phytosanitary measures followed European organic farming regulations. During the growing seasons, fertigation was done at modest doses providing around 16 kg ha−<sup>1</sup> N and 32 kg ha−<sup>1</sup> K2O (VIT-ORG) for all systems alike. The fertilization was meant to avoid K lack during fruit ripening, keeping the N supply at a minimum level (i.e., the amount of N contained in the NK fertilizer) in order to avoid masking the e ffects of treatments on N availability. The fertigation was practiced twice each year (when at least 70% of plants in all the plots reached the fruit set stage and two weeks later) with a single irrigation intervention early in the morning. Plots were 10 m × 6 m and 10 m × 5 m wide, respectively, in 2016 and 2017 and were distributed in a completely randomized block design over di fferent fields each year. The cover crop at killing dates yielded 2.3 (SD = 0.98) and 3.5 t ha−<sup>1</sup> (SD = 1.6) of dry biomass and had a N yield of 49.1 and 75.9 kg ha−<sup>1</sup> in 2016 and 2017, respectively. The soil was a sandy loam in 2015–2016 and a sandy clay loam in 2016–2017. Soil characteristics in each experimental site/year are detailed in Table 1. Weather conditions reported for the last 25 years and during the experiment are also presented in Figure 1.

**Table 1.** Soil characteristics of the fields where the experiments were carried out.


The Kjeldahl method was used for total N determination, the Walkley–Black method for soil organic matter (SOM), and the Olsen P test for soil available phosphorus (P) determination. EC = electrical conductivity.

**Figure 1.** Total monthly precipitations (mm) and the average maximum and minimum temperatures

( ◦C) during 2015 and 2017 compared to the multiannual average precipitations and temperatures (1993–2017) in San Piero a Grado, Pisa.

#### *2.2. Field Samplings and Measurements*

Tomato fruits were harvested from 12 plants of the central row of each plot through the season. The cumulative number of discarded (i.e., diseased, rotten, damaged), green and marketable tomatoes, and their corresponding fresh weights were recorded. Total yield as the sum of the fresh weights of all categories was therefore determined in order to estimate the potential cumulative yield of tomato. The dry matter (DM) content of fruits was obtained by oven-drying a sample at 60 ◦C until a constant weight was obtained. At the end of the harvest period, tomato residues and weeds were simultaneously collected over two areas of 1 m<sup>2</sup> in each plot. For crop residues, plants were excavated at depths of 25–30 cm and shoots and roots were separated after cleaning roots from soil residues. Plant parts were then oven-dried at 60 ◦C for dry matter and N content determination [19]. Tomato total N uptake (kg N ha−1) was calculated as:

$$\mathbf{N}\text{ uptake} = (\mathbf{a} \times \mathbf{b}) + (\mathbf{c} \times \mathbf{d}) + (\mathbf{e} \times \mathbf{f})\tag{1}$$

where "a" is tomato yield (kg ha−<sup>1</sup> of DM), "b" is the N concentration of marketable tomato fruits (g 100 g<sup>−</sup><sup>1</sup> of DM), "c" is the tomato shoot yield (kg ha−<sup>1</sup> of DM), and "d" is the N concentration of tomato shoot (g 100 g<sup>−</sup><sup>1</sup> of DM), "e" is the root yield (kg ha−<sup>1</sup> of DM), and "f" is the N concentration of tomato root.

To assess the dynamic status of nitrogen in the soil, the nitrate content was determined [20] every 10–20 days at a depth of 30 cm on a composite sample (2 samples) from each plot, starting at transplantation and continuing during the season. A hand-held electronic cone-tipped penetrometer (Spectrum Field Scout SC-900, Spectrum Technologies Inc., Plainfield, IL, USA) was used to measure soil resistance (KPa) on three di fferent locations in each plot at harvest across a 45 cm soil depth.
