Impact of Different Agroecological Practices for Weed Management on Weeds and Crops Development

Abstract

Cover crops and mulches are widely used techniques for limiting weeds and pests’ effects on crops. This study compared six practices over two growing seasons in two organic farms in Cuneo province, North-West Italy: two bio-based biodegradable mulch sheets (BM01 and BM02), dead mulch (hazelnut shells), living mulch (Trifolium repens L.), mechanical control, and an untreated control. Spring crops included Lactuca sativa L. var. capitata, Allium cepa L. cv. ‘Tropea’, and Brassica oleracea L. var. italica, while autumn crops were Lactuca sativa L. var. capitata, Allium fistulosum L., and Brassica oleracea L. var. italica. Weed infestation was evaluated through density (n/m²), biomass (g/m²), and diversity (Shannon Index), alongside crop yield and quality. Biodegradable mulch sheets provided the greatest weed suppression, followed by hazelnut shells, while living mulch and untreated control showed the highest weed pressure. Crop yield varied significantly among practices and species: BM01 and BM02 resulted in the highest yields, while living mulch consistently produced the lowest. Lettuce displayed the best quality across both farms, whereas onion quality varied by site. The highest quality scores were observed under biodegradable mulches and mechanical control, while living mulch and untreated control yielded the poorest results. Overall, biodegradable mulches emerged as the most effective balance between weed suppression, crop yield, and quality in organic systems.

1. Introduction

Weeds are among the most important biotic constraints in agriculture, causing an estimated 43% reduction in global crop yields [1]. The intensive reliance on synthetic herbicides is increasingly questioned due to environmental impacts, biodiversity loss, soil contamination, and risks to food security, while continuous use has also led to widespread cases of herbicide resistance [2]. Weed management thus remains a major challenge, particularly in organic systems, where the absence of chemical solutions requires the integration of physical, biological, and cultural tactics [3,4]. Weeds compete with crops for essential resources and can release allelochemicals that impair germination and growth, while their high seed production ensures future infestations [5]. For these reasons, developing effective agroecological strategies, based on indirect methods (competitive varieties, intercropping, localized fertilization, and irrigation) and preventive measures (clean seeds, crop rotation, and soil preparation), is crucial to minimize yield losses [6,7].

In organic systems, direct weed management methods include mechanical control, thermal weeding (pyro-weeding), and mulching. Among these, mulching is particularly effective because it suppresses weeds while also providing multiple agronomic and ecological benefits. In fact, mulch contributes to soil moisture conservation, temperature regulation, erosion prevention, and nutrient cycling, while simultaneously improving soil biological fertility [8,9,10,11]. Weed suppression occurs through several mechanisms, such as limiting light and moisture availability, acting as a physical barrier to emergence, or releasing allelopathic compounds [12].

Mulching materials can take different forms. Organic residues such as straw or nutshells not only create a protective soil cover but also recycle agricultural by-products, improve soil structure, and modify the microclimate to favor crop growth [13,14]. Biodegradable plastic sheets, produced from renewable resources, have been shown to provide strong and uniform weed control while reducing evapotranspiration losses. Living mulches, such as Trifolium repens, represent another strategy: they ensure permanent soil cover, contribute nitrogen fixation, and promote the activity of beneficial organisms [15,16,17,18,19]. However, their success depends on correct establishment and management, as they may compete with crops for light, water, and nutrients. Closely related are cover crops, which provide similar ecosystem services, including nutrient cycling, biodiversity enhancement, and pest regulation, but their introduction into rotations requires careful timing to avoid unintended competition [20,21,22,23,24].

Thus, while mulching and cover crops are essential tools for agroecological weed management, their effectiveness and trade-offs vary depending on the material used, crop species, and environmental context. Comparative field evaluations are therefore necessary to assess how different approaches perform in terms of weed suppression, crop productivity, and sustainability of the agroecosystem.

Another widely practiced mulching technique involves the use of plastic sheets, which are effective in improving yield and quality while reducing soil moisture loss. Currently, petroleum-based plastic films are applied on more than 80,000 km² of agricultural land worldwide, with an annual consumption of about 4.6 million tons [25]. However, their extensive use raises serious sustainability concerns, and the replacement of non-renewable plastics with biodegradable alternatives is increasingly required. New-generation films, produced from starch, vegetable oils, or cellulose by-products, are designed to be fully biodegradable and can be incorporated into the soil after use [11,26]. These solutions reflect a circular approach to mulching and respond to growing environmental concerns, particularly regarding climate change and resource depletion.

In this context, the present study compared six agroecological practices for weed management: two biodegradable mulch sheets, mechanical control, hazelnut shell mulch, clover (Trifolium repens) living mulch, and an untreated control. Trials were conducted on lettuce (Lactuca sativa L. var. capitata), onion (Allium cepa L. cv. ‘Tropea’ and Allium. fistulosum L.), and broccoli (Brassica oleracea L. var. italica). These crops were chosen because they are widely cultivated in organic systems, have economic relevance, and represent different functional groups (leafy, bulb, and inflorescence vegetables). Their contrasting growth characteristics and sensitivity to weed competition make them suitable for evaluating the performance of different agroecological weed management practices.

2. Materials and Methods

2.1. Experimental Design

2.1.1. Experimental Setup

A field experiment was carried out for two consecutive growing seasons (spring/summer 2021 and autumn/winter 2021/2022) in two organic soils of the Cuneo province: the garden of the University of Gastronomic Sciences (Pollenzo, Cuneo, North-East Italy) (Lat. 44°42′ N; Long. 07°51′ E) and the Orto del Pian Bosco (Fossano, Italy) (Lat. 44°32′ N; Long. 07°43′ E). Since the main differences between the two experimental sites are related to their soil characteristics, in the following sections we refer to them as “Soil 1” (Pollenzo) and “Soil 2” (Fossano) rather than “farms”.

Table 1. Site and soil characteristics.

Site (Alias)	Location (Lat, Lon)	Texture	pH	Total N (‰)	P2O5 (mg·kg−1)	K2O (meq·100 g−1)	Organic Matter (OM) (%)	Mean Temperature/Precipitation (SS21)	Mean Temperature/Precipitation (AW21/22)
Pollenzo (Soil 1)	44°42′ N; 07°51′ E	Sandy loam (12% clay, 53% sand, 35% silt)	7.1	2.7	74.6	0.31	3.5	19.9 °C/1.3 mm	10.1 °C/1.0 mm
Fossano (Soil 2)	44°32′ N; 07°43′ E	Clay loam (30% clay, 40% sand, 30% silt)	7.4 (sub-alkaline)	1.4	145.0	0.35	3.0	18.7 °C/1.7 mm	11.5 °C/1.4 mm

P₂O₅ (mg·kg⁻¹) assimilable; K₂O (meq·100 g⁻¹) exchangeable; SS21: spring/summer 2021; AW21/22: autumn/winter 2021–2022.

summarizes all the climatic and soil characteristics of the two sites analyzed in our study. Climatic data were obtained from ARPA Piemonte (Agenzia Regionale per la Protezione Ambientale, Piemonte). The preceding crop was tomatoes in Pollenzo and chili in Orto Pian Bosco. The same six agroecological weed management (AWM) practices were compared at both sites, i.e., two different bio-based biodegradable mulch sheets (BM01 and BM02), dead mulch with hazelnut shells (DM), living mulch with dwarf clover Trifolium repens, L. (LM), mechanical control (MC), and an untreated control. The experimental design is a split-split-plot with randomized block, where main factors are season, soil and sampling date, while AWM is the first sub-plot (six randomized rows, 27 m × 1 m each) and crop species is the sub-sub-plot (3 randomized crops, 9 m × 1 m each). AWM treatments were randomized within each agroecosystem in the first season, and the same layout was maintained in the second season to ensure comparability across trials. Crops transplanted during the first trial, which took place from April to July 2021, were Lactuca sativa var. capitata L., Allium cepa L. cv. ‘Tropea’ and Brassica oleracea var. italica, while in the second trial, which took place from September to March 2022, Lactuca sativa L. var capitata, Allium fistulosum L., and Brassica oleracea var. italica were used. The soil was harrowed before transplanting the crops and no fertilization was applied. In both trials, per each AWM treatment, the plant density was 12 p/m⁻² in three rows for each bed at 30 cm distance for Lactuca, 40 p/m⁻² in four rows for each bed at 10 cm for Allium, and 4.5 in two rows for each bed at 45 cm distance for Brassica (

Table 2. Crops, plant density, and spacing.

Crop	Taxon	Density (Plants·m−2)	Rows × Spacing
Lettuce	Lactuca sativa L. var. capitata	12	3 rows; 30 cm
Onion	Allium cepa L. cv. ‘Tropea’/Allium fistulosum L.	40	4 rows; 10 cm
Broccoli	Brassica oleracea L. var. italica	4.5	2 rows; 45 cm

). Crops were transplanted on 27 April and 9 September 2021.

2.1.2. Weed Management Treatments

For each of the six agroecological weed management (AWM) practices described above, specific materials and management practices were adopted (

Table 3. Agroecological weed management (AWM) treatments (applied at both sites—Pollenzo and Fossano).

Code	Practice (Type)	Material/Source	Key Operational Details
BM01	Biodegradable mulch sheet	Benchmark starch-based biopolymer (Novamont S.p.A.)	Laid pre-transplant; black film; left in field after use
BM02	Biodegradable mulch sheet	Starch-based biopolyester with bio-azelaic acid and bio-BDO blends (Novamont; PRIME project)	As BM01
DM	Dead mulch (hazelnut shells)	On-farm by-product	Uniform layer ≈ 3 cm covering soil
LM	Living mulch	Trifolium repens L., cv. ‘Rozeta’	Broadcast seeded by hand 25 g·m−2; 2 weeks pre-transplant
MC	Mechanical control	Hoeing + rotary harrowing	In-row hand hoeing; inter-row rotary harrowing (2×/week summer; 1×/month winter)
UC	Untreated control	—	No weed management

). The bio-based mulch sheets are produced from starch-based plastics. Starch-based plastics are blends of starch with one or more polymers (bio-based polyesters). These are obtained through a two-stage polymerization process based on transesterification and polycondensation using diacids and diols as monomers. The BM01 is a benchmark provided by Novamont S.p.A and obtained from virgin feedstock (i.e., vegetable oils and glucose). The BM02 developed and provided by Novamont S.p.A. (Novara, Italy) within the context of the PRIME research project (www.novamont.com/prime), on the other hand, is characterized by renewable diacid which is represented by azelaic acid produced from blends of virgin and/or waste vegetable oils (from food supply chains), while the diol is represented by 1,4 bio-BDO also obtained from blends of glucose and waste sugars from the food supply chain. The hazelnut shells came from Orto Pian Bosco farms and the thickness of dead mulch was 3 cm in each plot, enough to properly cover the soil. The Trifolium repens L., cv. ‘Rozeta’ was broadcast seeded by hand at 25 g·m⁻² two weeks before the crops transplanting, covering seeds by hand harrowing. Weeds growing in MC plots were controlled with hand hoeing and between the strip by rotary harrowing whenever needed, that was twice a week in the hot season and once per month in winter.

2.2. Evaluation of Weed Infestation

Collected data were number and biomass of weeds per each species, sampling three 0.5 × 0.5 m squares in each sub-sub-plots (replication). This allowed us to calculate weed density (n·m⁻²), weed biomass (g·m⁻²), and the Shannon Index (see data analysis). Botanical species were determined according to Pignatti [27] and the nomenclature reported according to the Portal to the flora of Italy (Portal to the Flora of Italy. Available online: https://dryades.units.it/floritaly/ (accessed on 24 June 2022)). Measurements were carried out weekly from transplanting until the end of harvest. To measure dry weight of weeds, fresh samples were maintained in dryer ATACAMA PRO Tre Spade (Turin, Italy) at 60 °C until a constant weight.

2.3. Evaluation of Crop Yield and Quality Observations

Marketable crop yield for each harvested crop was assessed (gr·m⁻²). The quality surveys included an assessment of the harvest by visual analysis of the crops. In accordance with Reg. (EU) 543/2011, which classifies marketing categories in terms of fruit and vegetable quality into “0” extra superior, “1” good quality, “2” minimal quality, and “3” unmarketable [28], each crop was assigned a value according to its appearance and any damage suffered. For this work, categories 0 and 1 were not considered as it is highly unlikely to assign these values in the organic sector, so category 2 was divided into “2a” major quality and “2b” minor quality.

2.4. Data Analysis

From the number of weeds per species (n·m⁻²), we derived the biodiversity Shannon Index defined as $H_j=-{\sum }_i{p}_{ij}\ln \left(p_{ij}\right)$ [29], where

Hj: The Shannon Index for the specific community or area being analyzed (in this case, for weeds per square meter).

p_ij: The proportion of individuals belonging to species i relative to the total number of individuals in the community j. It is calculated as p_ij = n_ij/N_j where n_ij is the number of individuals of species i in community j, and N_j is the total number of individuals of all species in community j.

Σi: The summation across all species in the community.

ln(p_ij): The natural logarithm (base e) of the proportion p_ij.

For both the analysis of weeds and yield, the data obtained from the assessments of the number of weeds per species (n × m⁻²), their biomass (g/m²), weed diversity (Shannon Index), harvest yields (g × m⁻²), and quality (2a and 2b), were independently submitted to ANOVA as a mixed-effect model where the variables “Year”, “Soil”, “Dates within Year” and “Species” are random-effects factors, while “AWM” is considered as a fixed-effect factor (see the attached). An example of the number of weeds is where n = repetitions (3), s = season (2), f = soil (2), d(s) = dates within season (9 for the first year and 13 for the second), m = agroecological weed management (6) and sp = species (3). The dates of the measurements is not considered as repetitive observations, but as an orthogonal levels factor, to reduce the complexity of the ANOVA model (Figure 1) and to analyze the direct interactions with the other factors, especially with Mulching. For variables related to yield and quality, the number of replicates per treatment combination varied depending on the species grown and the trial design, with a minimum of 5 and a maximum of 50. Although the actual number of observations reached up to 134 in some treatments, the harmonic mean of replicates across treatments was 16.97. This approach ensured a balanced statistical analysis despite differences in replication. All analyses were conducted using the IBM SPSS Statistics software, 28.0.1 version (IBM Corp., Armonk, NY, USA).

3. Results

3.1. Weed Density

A total of 46 weed taxa were identified across all plots in the two seasons, as shown in

Table 4. List of weed species identified and their botanical families.

Scientific Name	Botanical Families
Abutilon theophrasti Medik.	Malvaceae
Amaranthus retroflexus L.	Amaranthaceae
Blitum bonus-henricus (L.) Rchb.	Amaranthaceae
Calepina irregularis (Asso) Thell.	Brassicaceae
Capsella bursa-pastoris (L.) Medik.	Brassicaceae
Centaurea cyanus L.	Asteraceae
Cerastium glomeratum Thuill.	Caryophyllaceae
Chenopodium album L.	Amaranthaceae
Convolvulus arvensis L.	Convolvulaceae
Crepis sp.	Asteraceae
Cyperus sp.	Cyperaceae
Erechtites hieraciifolius (L.) Raf. ex DC.	Asteraceae
Galinsoga parviflora Cav.	Asteraceae
Geranium dissectum L.	Geraniaceae
Geranium molle L.	Geraniaceae
Lamium sp.	Lamiaceae
Lathyrus tuberosus L.	Fabaceae
Lolium multiflorum Lam.	Poaceae
Lotus pedunculatus Cav.	Fabaceae
Malva sylvestris L.	Malvaceae
Marrubium vulgare L.	Lamiaceae
Matricaria chamomilla L.	Asteraceae
Medicago lupulina L.	Fabaceae
Medicago sativa L.	Fabaceae
Mentha pulegium L.	Lamiaceae
Nasturtium officinale W.T. Aiton	Brassicaceae
Oxalis sp.	Oxalidaceae
Parietaria officinalis L.	Urticaceae
Plantago major L.	Plantaginaceae
Polygonum aviculare L.	Polygonaceae
Portulaca oleracea L.	Portulacaceae
Potentilla reptans L.	Rosaceae
Ranunculus repens L.	Ranunculaceae
Raphanus raphanistrum L.	Brassicaceae
Rorippa sylvestris (L.) Besser	Brassicaceae
Rumex sp.	Polygonaceae
Senecio vulgaris L.	Asteraceae
Sonchus sp.	Asteraceae
Stellaria media (L.) Vill.	Caryophyllaceae
Taraxacum sect. Taraxacum F.H. Wigg	Asteraceae
Trifolium sp.	Fabaceae
Valeriana locusta L.	Caprifoliaceae
Veronica anagallis-aquatica L.	Plantaginaceae
Veronica persica Poir.	Plantaginaceae
Vicia cracca L.	Fabaceae
Viola palustris L.	Violaceae

N.b. according to Plants of the Word Online.

. Out of the 20 botanical families represented, the most common were Asteraceae (n = 8 species), Fabaceae (n = 6), and Brassicaceae (n = 5). A few species accounted for the majority of weed abundance across treatments. In particular, Portulaca oleracea L. and Stellaria media (L.) Vill. were consistently among the most abundant species, while Lamium spp. and Veronica persica Poir. were also frequent in several treatments. By contrast, some taxa such as Potentilla reptans L. and Viola palustris L. occurred only sporadically. It should be noted that Cynodon dactylon (L.) Pers. was not included in the species table because it was not part of the main categorization used in this study; however, it was frequently observed during field surveys and represented one of the most abundant weeds in the experimental sites, reflecting its widespread distribution in the local agroecosystems.

The results of the ANOVA (

Table 5. Effect of variables on weed density.

Source of Variance	SS	d.f.	MS	F	p	Sign.
Season “S”	155	1	155	0.004	0.948	n.s.
Soil “So”	2,218,925	1	2,218,925	62.3	<0.0001	**
Interaction S × So	11,865,99	1	1,186,599	33.3	<0.0001	**
Date within Season “D(S)”	1,193,593	20	59,680	1.7	0.143	n.s.
Error a	605,684	17	35,628
AWM (M)	17,054,487	5	3,410,897	27.2	0.001	**
Interaction S × AWM	625,900	5	125,180	2.5	0.035	*
Interaction So × AWM	4,215,985	5	843,197	16.9	<0.0001	**
Interaction AWM × D(S)	8,362,792	100	83,628	1.7	0.006	**
Error b	4,693,483	94	49,931
Species within Season “Sp(S)”	222,706	4	55,676	2.0	0.155	n.s.
Interaction So × Sp(S)	55,952	2	27,976	4.2	0.015	*
Interaction M × Sp(S)	85,836	13	6603	1.001	0.448	n.s.
Error c	8,829,586	1338	6599
Total	49,351,682	1630

S = season; So = soil; AWM (M) = agroecological weed management practice; D(S) = date within season; Sp(S) = species within season. n.s. = p > 0.05; * = p ≤ 0.05; ** = p ≤ 0.01.

), applied to analyze which variables might affect weed density, revealed several significant differences, mainly related to the soil (p < 0.0001) and AWM (p = 0.0012). Additionally, several interactions were found to be statistically significant, including “Season × Soil” (p < 0.0001), “Season × AWM” (p = 0.035), “Soil × AWM” (p < 0.0001), “AWM × Date within Season” (p = 0.006), and “Soil × Species” (p = 0.015). In particular, through analyzing the effect of the AWM practices (Figure 2), it was observed that the practices providing the highest level of weed control were the two biodegradable mulches, mechanical control, and hazelnut shell mulching, with no significant differences among them. Conversely, LM showed significantly different results, and as expected, the untreated control exhibited the highest density of weeds.

Bars represent mean values ± SE. Different letters above bars indicate significant differences among treatments according to Tukey’s HSD test (p < 0.05). Data are pooled across both soils and growing seasons. BM01 = biodegradable mulch sheet 01; BM02 = biodegradable mulch sheet 02; DM = dead mulch (hazelnut shells); MC = mechanical control; LM = living mulch; Control = untreated control.

However, analyzing the significant interaction between “Season × AWM” (p = 0.035), as shown in Figure 3, while it is not possible to compare two different seasons statistically, it can be observed that weed density remained unchanged across seasons for BM01, BM02, and LM. In contrast, for MC, DM, and the untreated control, higher weed density values were observed during Season 1 compared to Season 2.

Regarding the significant interaction between “Soil and AWM,” as shown in Figure 4, the practices exhibited the same trends in weed control across both soils, except for LM and the untreated control, which were less effective in Soil 1 compared to Soil 2.

3.2. Weed Biomass and Diversity

The results of the ANOVA conducted to analyze the main effects of the considered variables on biomass values revealed a single significant difference (p < 0.0001) related to the interaction between “Season × AWM.” Specifically, analyzing this interaction, as shown in Figure 5, it was observed that, consistent with the data obtained from the weed density analysis, weed density remained unchanged across the two seasons for BM01, BM02, and LM. In contrast, for MC, DM, and the untreated control, higher weed density values were recorded during Season 1 compared to Season 2. The results of the ANOVA (

Table 6. Effect of variables on weed diversity.

Source of Variance	SS	d.f.	MS	F	p	Sign.
Season “S”	0.060	1	0.060	0.625	0.440	n.s.
Soil “So”	0.323	1	0.323	3.357	0.084	n.s.
Interaction S × So	3.852	1	3.852	40.03	<0.0001	**
Date within Season “D(S)”	16.32	20	0.816	8.480	<0.0001	**
Error a	1.636	17	0.096
AWM “M”	30.91	5	6.183	5.366	0.044	*
Interaction S × M	5.761	5	1.152	6.246	<0.0001	**
Interaction So × M	4.425	5	0.885	4.797	0.001	**
Interaction M × D(S)	22.14	100	0.221	1.200	0.186	n.s.
Error b	17.34	94	0.184
Species within Season “Sp(S)”	0.790	4	0.197	0.466	0.759	n.s.
Interaction So × Sp(S)	0.359	2	0.180	1.609	0.200	n.s.
Interaction M × Sp(S)	5.501	13	0.423	3.789	<0.0001	**
Error c	149.4	1338	0.112
Total	259	1630

S = season; So = soil; AWM (M) = agroecological weed management practice; D(S) = date within season; Sp(S) = species within season. n.s. = p > 0.05; * = p ≤ 0.05; ** = p ≤ 0.01.

) applied to analyze the main factors potentially affecting Shannon Diversity Index values revealed significant differences, particularly based on the date of the trial (p < 0.0001) and AWM (p = 0.0445). Specifically, analyzing the effect of AWM practices (Figure 5), it is evident that, as expected, the untreated control preserved the highest biodiversity, although it was not significantly different from LM and MC practices. These latter two practices were also significantly similar to the DM practice. Additionally, the figure shows a significant similarity between the DM and BM02 practices. Finally, BM01 appeared to be the least effective practice in terms of biodiversity, although it was not significantly different from BM02. Significant interactions were also observed, including “Soil × Season” (p < 0.0001), “Season × AWM” (p < 0.0001), “Soil × AWM” (p < 0.0006), and “AWM × Species” (p < 0.0001).

Furthermore, examining the significant interaction between “Species × AWM” (Figure 6) and between “Species × AWM × Season” (Figure 7), it was found that, as expected, the untreated control exhibited the highest Shannon Index for all three crops. However, the practices with clover and mechanical control also displayed high Shannon Index values for all three species. For Brassica oleracea var. italica and Lactuca sativa var capitata L., the biodegradable mulch 01 and the experimental biodegradable mulch 02 showed lower diversity levels compared to the other practices. Analyzing the significant interaction between “Season and AWM” (p < 0.0001), it was observed that Season 2 was characterized by higher biodiversity levels across all practices, except for BM02, which showed overlapping trends for both seasons.

3.3. Crop Yield

The ANOVA analysis applied to the data obtained from the evaluation of crop yields in the two experimental trials revealed significant differences, particularly for the factors AWM and the interaction “Season × Species” (p = 0.023, p < 0.0001, respectively;

Table 7. Effect of variables on crop yield.

Source of Variance	SS	d.f.	MS	F	p	Sign.
Season “S”	22,767.8722	1	22,767.8722	65.1468	0.078	n.s.
Soil “So”	2682.5011	1	2682.5011	7.6756	0.221	n.s.
Interaction S × So	349.4856	1	349.4856	0.3971	0.556	n.s.
AWM “M”	32,510.2035	5	6502.0407	7.3878	0.023	*
Interaction S × M	12,503.9764	5	2500.7953	2.8415	0.138	n.s.
Interaction So × M	8841.9046	5	1768.3809	2.0093	0.231	n.s.
Error a	4400.5118	5	880.1024
Species “Sp”	114,268.1427	2	57,134.0713	2.5808	0.092	n.s.
Interaction S × Sp	44,276.5908	2	22,138.2954	24.9867	<0.0001	**
Interaction So × Sp	5274.7855	2	2637.3927	2.9767	0.066	n.s.
Interaction M × Sp	7636.9069	10	763.6907	0.8619	0.577	n.s.
Error b	26,580.1078	30	886.0036
Residue	314,454.2030	4963	63.3597
Total	596,547.1918	5032

S = season; So = soil; AWM (M) = agroecological weed management practice; Sp = species. n.s. = p > 0.05; * = p ≤ 0.05; ** = p ≤ 0.01.

). The graph presented in Figure 8 shows the mean values of crop yields, averaged across all three tested species, highlighting that the only significant difference is between the BM01 and LM practices, with the latter displaying the lowest values. The other practices were not significantly different from either BM01 or LM. Furthermore, for Allium cepa L. var. ‘Tropea’, the highest productivity was observed in Season 1, followed by a significant decrease in Season 2. For Brassica oleracea L., yields were consistently low and comparable across both seasons. Finally, for Lactuca Sativa var capitata L., the highest yields were recorded in both seasons, with higher values observed in Season 1.

3.4. Quality of Yield

Table 8. Effect of variables on quality yield.

Source of Variance	SS	d.f.	MS	F	p	Sign.
Season “S”	13.8130	1	13.8130	0.4137	0.636	n.s.
Soil “So”	66.1287	1	66.1287	1.9808	0.393	n.s.
Interaction S × So	33.3854	1	33.3854	85.0217	0.0003	**
AWM “M”	35.1993	5	7.0399	17.9282	0.003	**
Interaction S × M	8.9338	5	1.7868	4.5503	0.061	n.s.
Interaction So × M	4.3007	5	0.8601	2.1905	0.205	n.s.
Error a	1.9633	5	0.3927
Species “Sp”	240.7656	2	120.3828	4.0541	0.028	*
Interaction S × Sp	10.9625	2	5.4812	1.8762	0.171	n.s.
Interaction So × Sp	59.3879	2	29.6939	10.1641	0.0004	**
Interaction M × Sp	15.5768	10	1.5577	0.5332	0.853	n.s.
Error b	87.6435	30	2.9215
Residue	414.9726	4963	0.0836
Total	993.0331	5032

S = season; So = soil; AWM (M) = agroecological weed management practice; Sp = species. n.s. = p > 0.05; * = p ≤ 0.05; ** = p ≤ 0.01.

applies to the data on crop quality evaluation, and revealed significant differences, particularly for the factors AWM and crop species (p = 0.003, and p = 0.028, respectively). Additionally, the interactions “Season × Soil” (p = 0.0003) and “Soil × Species” (p = 0.0004) were found to be statistically significant. The graph presented in Figure 9 shows the mean values from both trials, representing the quality of the harvested crops. Specifically, Lactuca Sativa var. Capitata L. was characterized by the highest quality levels across both soils. Following this, cabbage showed comparable quality trends between the two soils. In contrast, for Allium cepa L. var. ‘Tropea’, the quality of the harvest varied between the two farms, with Soil 1 yielding lower quality compared to Soil 2. The highest quality scores were observed for the biodegradable mulch sheets and mechanical control (Category 2a). The hazelnut shell mulching practice achieved a quality score that was not statistically different from BM01, BM02, and MC, nor from LM and the untreated control. Among these, LM and the untreated control were the practices that provided the lowest crop quality.

4. Discussion

Spontaneous flora is a crucial element of agroecosystem but the strategies for crop yield are highly dependent on weed management [30]. In organic farming, weeds can be a limitation to crop yield and quality, due to competition for nutrients, water, light, air, and space [5,31]. The mechanical control adopted to manage weeds and reduce the weed seed bank present in the different soil layers can then affect its physical, chemical, and biological properties with possible repercussions on the soil health [32]. It is, therefore, crucial to choose the most appropriate practices in managing weeds, not harming the agroecosystems or impacting crop yield. Indeed, this study aimed at comparing agroecological weed management practices allowed in organic farming, and the results showed that the different treatments applied significantly affected weed density and diversity as well as crop’s yield quantity and quality.

In particular, the present research showed that among the analyzed practices, biodegradable mulch sheets were found to result in better weed control in both years. This may be due to their ability to prevent the passage of light, which is essential for photosynthesis and necessary for weed growth, as shown by [33]. This more efficient weed management could have also contributed to higher yields and quality for all the crops analyzed. This can be attributed to the fact that biodegradable mulch sheets retain water in the soil, reducing evapotranspiration losses. As a result, they contribute to increased aeration and nutrient uptake by the plants [34]. Favorable moisture and temperature levels under the mulch imply an increase in stable soil aggregates, which counteracts in establishing favorable growth conditions for plant roots, also increasing root secretion, and results in increased nutrient availability for microorganisms, thus ensuring a stable agroecosystem in the soil [35,36]. Despite this efficacy against weeds, however, such control of biodegradable sheets has resulted in lower weed diversity compared to other practices.

Among the other agricultural practices examined, mechanical control was found to be efficient in controlling weeds and resulting in good crop yields, while also in high weed diversity. Crop performance and soil properties can be affected by tillage practices and reduced mechanical control, such as what was applied in this work, which may have contributed to increased soil performance by improving moisture infiltration and water use efficiency [37,38]. The present work hypothesized that using plant wastes as mulch could be a more viable alternative for weed control than even conventional methods, as found in the study by El-Metwally et al. [39] in which the use of vegetable waste as a ground cover significantly reduced weed biomass and increased the potential yield and quality of sugar beet, or also in the studies by Chang et al. [40] and Sinkevičienė et al. [41] in which a straw mulch was found to inhibit weed growth, compared to non-mulched fields. In this work, a mulch with hazelnut shells was sufficiently effective against weed growth as it was able to block light from reaching the soil surface by reducing germination and suppressing weeds [42]. Arguably, more nut shells, and thus more homogenous soil cover, could have been more effective, as it is possible that sunlight may have been able to penetrate the soil anyway, resulting in more competition between crops and weeds for soil minerals, soil moisture, and CO₂, which becomes more available to weeds [39]. In fact, crops may have been affected by competition with weeds, decreasing crop productivity and quality of yield. These results are partially in disagreement with the study by El-Metwally et al. [39] in which a peanut mulch was found to have significantly improved yield and crop quality. Mulching with peanut shells was not found to be a particularly effective practice in terms of soil biodiversity, although decomposition of organic mulch by soil microorganisms should produce high soil organic matter, increase soil biodiversity, and ensure ecosystem functions [38]. In addition to the overall density of weeds, the species composition revealed further information. Most of the abundant taxa were edible weeds traditionally consumed in Piedmont, such as Portulaca oleracea L. and Stellaria media (L.) Vill. In contrast, Cynodon dactylon (L.) Pers., although not listed in the species table, was frequently observed and represents one of the most problematic inedible weeds in the area. In particular, this species was strongly suppressed in the hazelnut shell mulch treatment, while it remained abundant in other systems. This suggests that mulching with dead plant residues can contribute not only to reducing overall weed pressure, but also to controlling highly competitive perennial species such as Cynodon dactylon (L.) Pers. However, it should be noted that the large-scale applicability of hazelnut shells as mulch is strongly dependent on their local availability as an agricultural by-product, and their use may therefore be limited to areas where hazelnut cultivation and processing are widespread. In this work, living mulching with clover was proved as not being particularly effective in managing weeds, although a reduction in weed biomass was found in the winter period compared to the trial conducted in the spring/summer period. The results obtained in this work, however, differ from the results obtained by Fracchiolla et al. [43] in which they found that clover was able to control weeds and compete with weed vegetation in the same way as biodegradable film, making it a viable alternative. In terms of productivity and crop quality, clover cover was found to have a negative impact in this study, which was also confirmed by Radicetti et al. [44], who found that the use of clover in winter wheat negatively affected growth and yield whilst reducing weed density, and den Hollander et al. [45], who reported that clover, used in leek crops, resulted in reduced plant weight, as they were completely entangled in the clover canopy. It should be considered that living mulch competes with weeds for light, soil moisture, and nitrogen (N). Therefore, such competition can affect seedling development and inhibit seedling growth [46]. However, Fracchiolla et al. [43] showed the positive effects of living mulch and organic fertilization in broccoli rabe production, both in terms of pest management, crop yield, and quality. The different results may depend on the different ecophysiological characteristics of the different crops and, most importantly, on the duration of their critical period of competition with weeds [43]. In our trials, the Trifolium repens, that were supposed to be dwarf varieties, resulted in high growth anyhow, competing with crops. Moreover, the clover should have been seeded earlier for a more homogeneous establishment and a uniform canopy of the living mulch.

These results indicate that different agroecological practices influenced both the abundance and identity of the weed community. Biodegradable mulch sheets and mechanical control reduced the presence of aggressive weeds while allowing edible taxa to persist, whereas living mulch promoted greater weed biodiversity but was less effective at suppressing competition with crops. Such patterns may have implications for agroecosystem functioning, crop productivity balance, weed management, and the valorization of wild edible species.

One limitation of this study is that AWM treatments were not re-randomized between the two seasons. While this approach allowed for direct comparability of treatments across years, it may also have introduced some spatial bias. Nevertheless, the inclusion of ‘Season’ as a factor in the statistical model helped account for temporal variability and mitigate this potential effect.

Such patterns may have implications for agroecosystem functioning, crop productivity balance, weed management, and the valorization of wild edible species. However, although agroecological practices significantly reduced weed density and diversity, this did not consistently translate into higher crop yields. This discrepancy may be explained by crop-specific physiological constraints, seasonal effects, and environmental variability, which may override the benefits of weed suppression. For example, Brassica oleracea showed uniformly low yields irrespective of weed management, while Allium cepa was mainly influenced by seasonal conditions rather than weed pressure. Moreover, quality parameters appeared more responsive to agroecological practices, with biodegradable mulches and mechanical control improving crop quality despite not always enhancing yield. These findings suggest that weed suppression alone is not sufficient to guarantee yield increases, as productivity in organic systems depends on a broader set of agronomic and ecological factors.

Moreover, the different weed management practices were compared with an untreated treatment, in which no intervention, either direct or indirect, on the weeds or soil took place. Non-weed management resulted in lower yield and crop quality of cabbage and onion productions; however, lettuce crop yield on average over the two years was higher than in the others and soil biodiversity was higher. A zero approach can result in several beneficial effects, such as minimal soil erosion damage, reduced soil disturbance, and reduced soil evaporation, while mechanical control, especially when poorly managed, can result in hard soil structure thereby limiting root growth and hindering plant growth and nutrient and moisture extraction from deeper soil layers.

5. Conclusions

This study aimed at comparing the different agroecological weed management practices allowed in organic farming, and the results showed that the different treatments applied significantly affected weed quantity and diversity and horticultural crop yield quantity and quality. In particular, the use of bioplastic sheets provided effective weed control while achieving high yields in terms of productivity and quality. However, both sheets were less efficient in preserving weed biodiversity. Mechanical control achieved comparable results in terms of weed control and crop yield while preserving high weed biodiversity. Dead mulching was quite efficient according to the characteristics examined, although at lower levels. The highest values of weed number, biomass, and diversity were recorded for the clover living mulch and the untreated treatment for all three crops as well as the lower values for yield quantity and quality, probably due to wrong varieties of clover and not proper seeding period. Thus, this study demonstrated how different agroecological practices on weed management can effect the crop yield and agroecosystem. Future studies should explore the long-term performance of these practices under different environmental conditions and cropping systems, as well as their economic feasibility. Moreover, policy measures supporting the adoption of biodegradable mulches and diversified weed management strategies could play a crucial role in promoting sustainable and resilient organic farming systems.