Differences in Weed Taxa Community in a Young Apple Orchard (‘King Roat Red Delicious’ Cultivar) Depending on the Presence of Living Mulch and the Application of Two Nitrogen Fertilization Rates

Abstract

The objective of this study was to evaluate the impact of two nitrogen doses in combination with strong creeping fescue (Festuca rubra L. ssp. rubra Gaudin) and Chewing’s red fescue (Festuca rubra L. ssp. commutata Gaudin) used as living mulches on the weed community in an apple tree (Malus domestica Borkh.) orchard. The cover grasses were sown in the tree rows, and herbicide fallow served as the control. Grass living mulches effectively reduced the number and share of annual weed cover and limited the spread of perennial plants compared with herbicide fallow. Use of F. rubra L. subspecies did not favor the biodiversity of the orchard agroecosystem flora, due to the effective soil surface coverage by sod in the tree rows. Living mulch sod was characterized by lower variability in weed taxa compared with the abundant weed composition in the herbicide fallow, which also exhibited the highest number of weed taxa each year. Dominant species in the orchard across all treatments included Trifolium repens L. and Taraxacum spp. Doubling the nitrogen fertilization rate, while limiting the application area to the tree canopy, did not increase the perennial weed population in the living mulch sod. Both subspecies are useful as living mulch in a young apple orchard, but from the perspective of sod durability and weed control, strong creeping red fescue offers better prospects.

1. Introduction

The broad spectrum of action and low cost of glyphosate have led to its widespread use in agriculture [1]. However, modern approaches to crop cultivation recommend limiting chemical weed control methods in favor of non-chemical weed management. Research on methods to eliminate glyphosate in plant production was already very relevant at the beginning of the 21st century [2,3], when herbicide fallow was commonly used as orchard floor management in tree rows in commercial orchards [4]. Two decades later, this topic has not lost its importance [5,6], especially in the context of the increasingly questionable future use of glyphosate in plant cultivation [1,7]. One possible method is to use cover crops as living mulch, an alternative method of orchard floor management practice that replaces herbicide fallow [7,8,9], especially in organic orchards [10,11,12].

Selecting cover crops that will provide good soil surface coverage, suppress weeds, and do not compete excessively with the main crop is crucial when using living mulch [13]. Some species from the Poaceae and Fabaceae families, as well as creeping plants from other botanical families, are useful [14]. An example of such a species is red fescue (Festuca rubra L.), which belongs to the group of fine fescues characterized by narrow leaves and low input requirements [15,16]. Taxonomic varieties of this grass have varying degrees of underground stolon growth, which allows for the development of a very strong, compact turf. This species is drought-resistant and tolerates two weeks of waterlogging. F. rubra L. is widely available in Poland, and its economic importance is increasing [17]. Its two subspecies, F. rubra ssp. rubra Gaudin—characterized by well-developed rhizomes with a creeping growth habit—and F. rubra ssp. commutata Gaudin—with bunch-type growth habit—are characterized by the production of dense sod that covers the soil well [15,18,19]. This is due to the favorable growth rate of seedlings and root system development compared with other grass species [20]. In orchard studies, in rows of young trees and tractor alleys, F. rubra ssp. rubra Gaudin covered almost 100% of the surface already in the second year after sowing, just like Lolium perenne L.—a species with a very short germination period and rapid sod development [21]. The good development of the sod is confirmed by pot experiments, where especially the subspecies F. rubra ssp. rubra Gaudin was characterized by a higher dry biomass compared with F. rubra ssp. commutata Gaudin [22]. Field evaluation of this grass showed a tendency for better soil cover by strong creeping fescue [23]. Such features of the F. rubra L. subspecies encourage its use in the orchard as living mulch. Moreover, this species effectively limits the weed population in the orchard [21,22] compared with other frequently used grass species in the orchard, such as L. perenne L. or Poa pratensis L. [21]. Nevertheless, after several years of use, other grasses and dicotyledonous weed species are present in its sod [24].

In modern fruit growing, the key is to control the weed taxa population but not to completely eradicate it. Weeds are both harmful and beneficial to the biodiversity of the agricultural landscape [25,26]. Their low population can coexist with the main crop without significant losses in the quality of cultivated plants and their yield [7,27]. As some experiments have shown, weed management practices have a significant effect on the characteristics and number of weed species in the orchard [12,21,28,29]. Herbicide fallow is dominated by annual weeds, with a short growing season and rapid growth between herbicide applications. There are also perennial weed taxa that can survive under repeated agrochemical treatments [30,31]. The presence of living mulch sod limits the development of annual species and stabilizes the share of perennial species cover [21] over several years after the introduction of living mulch to the orchard [32]. The most frequently occurring weed species in various experiments with grass living mulch were Taraxacum spp., Trifolium repens L., Plantago spp., Malva sylvestris L., and Rumex obtusifolius L., as well as Elymus repens (L.) Gould and other perennial grasses [10,22,33,34]. These species prefer environments ranging from moderately rich in nitrogen up to nitrogen-rich conditions [35]. This means that the presence of living mulch and its weeds creates competition for the fruit tree [36,37]. Nitrogen is necessary to maintain the proper health status of a tree [38]. Standard nitrogen fertilization involves applying 20–80 kg N ha⁻¹ depending on soil type in apple tree cultivation [39]. In an orchard, potentially three plant populations—fruit trees, grass cover crops, and weeds—benefit from this macroelement. Applying higher doses may reduce competition from additional plants against the fruit tree. Tripling the basic nitrogen rate (60 kg ha⁻¹) in the apple orchard ensured a high nitrogen concentration in the leaves of apple trees grown in living mulch compared with other orchard floor management systems [40]. Weedy species will also benefit from the nitrogen fertilizer applied, with varying degrees of success. Nitrogen application may lead to changes in the ratio between living mulch and weed populations. From one perspective, nitrogen fertilization influenced the initial development of fine fescue sod but also the dynamic development of weeds, e.g., Poa annua L. invasion. Increasing the nitrogen dose from about 100 to about 120 kg N ha⁻¹ had no effect on the P. annua L. population in spring, but at the beginning of summer it significantly increased the share of this weed in the grass sod [36]. Increasing the nitrogen fertilization rate from about 50 to about 100 kg N ha⁻¹ significantly reduced the percentage of weed cover in hard fescue sod [41]. On the other hand, a high nitrogen rate, up to about 200 kg ha⁻¹, reduced the percentage of weed cover in F. rubra ssp. commutata Gaudin sod [37].

The aim of the study was to evaluate the impact of two nitrogen doses on plant community composition and biodiversity of weed taxa occurring in tree rows with living mulches composed of two subspecies of F. rubra L. in comparison with the use of herbicide fallow in a young apple orchard.

2. Materials and Methods

2.1. Orchard Location

The influence of grass living mulch, as an alternative floor management to herbicide fallow, and two doses of nitrogen fertilization on the weed species composition in a young apple orchard was estimated. The experiment was conducted at the Fruit Experimental Station of the Wrocław University of Environmental and Life Sciences (Poland), from 2021–2024. It was located in Samotwór (51°06′12″ N, 16°49′52″ E), near the city of Wrocław. Two-year-old apple trees (Malus domestica Borkh.) of the ‘King Roat Red Delicious’ cultivar, grafted on semi-dwarf M.26 rootstock, were planted with a spacing between rows of 3.5 m, and a 1.4 m distance between trees was used in a row.

2.2. The Experimental Factors

The experiment was established following a two-way randomized block design, with four replications. The first experimental factor was orchard floor management, in which the commonly used herbicide fallow in fruit tree rows was replaced with grass living mulch. Strong creeping red fescue F. rubra L. ssp. rubra Gaudin (hereafter, FRR) cultivar ‘Oaza’ and Chewing’s red fescue F. rubra L. ssp. commutata Gaudin (hereafter, FRC) cultivar ‘Nawojka’ were used as two living mulches. Each of the cover grasses was sown separately, on plots with an area of 7 m², marked in rows 1 m wide, in late summer 2021.

One month prior to grass seed sowing, the soil surface was prepared by chemical weed control using a mixture of two herbicides: glyphosate, 1.96 kg ha⁻¹ (Trustee Hi Activ, Barclay Chemicals, Dublin, Ireland), and 2-methyl-4-chlorophenoxyacetic acid (MCPA), 0.60 kg ha⁻¹ (Chwastox Extra 300 SL, Ciech, Nowa Sarzyna, Poland). Approximately five weeks later, the soil was cultivated manually using a hoe, and most of the remaining root systems of the perennial weeds were carefully removed by hand. The tree-row area was subsequently leveled with a rake. On 25 August 2021 grass seeds were manually broadcast, lightly incorporated into the soil with a rake, and the surface was then compacted using a hand roller. The seeds of the two investigated red fescues were obtained from Małopolska Hodowla Roślin HBP Sp z. o. o. (Kraków, Poland) and were sown at the seed rate of 50 kg ha⁻¹. In the living mulch plots, no chemicals were applied to the grass sod in the subsequent years of the experiment. Since spring 2022, the sod of the living mulches has been manually mowed using a string trimmer (Sthil FS 360 C-EM, Waiblingen, Germany), two or three times per growing season, depending on vegetation growth developments. In 2022, it was conducted in May, June, and August. In the two following years, 2023–2024, mowing was limited to two times: May and June, and April and July, respectively.

Herbicide fallow plots, as a control orchard floor management, were treated with a mixture of glyphosate (1.44 kg ha⁻¹) and MCPA (0.60 kg ha⁻¹). It was applied two times per vegetation season. Glyphosate was mainly applied in the form of the herbicide Singlif 360 SL (Agronomica, Hamburg, Germany); however, in two applications, Gallup Special 360 (Ciech, Nowa Sarzyna, Poland) and Dominator Clean 360 SL (Albaugh, Rače, Slovenia) were used as alternative glyphosate sources. MCPA was consistently applied as Chwastox Extra 300 SL (Ciech, Nowa Sarzyna, Poland). In accordance with current recommendations for commercial orchards [39], treatments were performed on the following dates: May and July, 2022; May and August, 2023; and April and July, 2024. Herbicide application was performed using a ‘Sano 2’ backpack sprayer (Agromet–Pilmet, Wrocław, Poland) with a capacity of 12 L, 500 kPa working pressure, and an output of 1.5 L per minute. Spraying was performed separately on each side of the apple tree row, at a working width of 0.5 m.

From spring 2022, a second experimental factor was introduced—nitrogen fertilization. Ammonium nitrate (34% N, Anvistar, Anvil S.A., Włocławek, Poland) was applied manually, separately stricte under each tree canopy, at a dose corresponding to 50 and 100 kg N ha⁻¹. The initial fertilizer dose of 50 kg N ha⁻¹ was determined based on current recommendations for commercial orchards [39]. Doubling the dose was intended to improve the growth conditions of the apple trees in grass cover crops. In subsequent years, nitrogen was applied once during the growing season in mid-May. No nitrogen fertilization was applied in the year of orchard establishment (2021).

2.3. Agrotechnical Methods in Experiment

The tree canopies were shaped like a slender spindle (conical shape). Tree pruning and protection were performed exactly as in a commercial orchard, in accordance with the recommendations of the Methodology of Integrated Apple Production [39]. In a dry year, 2022, experimental plots were emergency irrigated six times with an over-head irrigation system. It was conducted from June up to August. Average water rate was approximately 3 L tree⁻¹ (15 L per plot). In the case of two irrigations, the water rate was reduced to about 1.5 L tree⁻¹ (7.5 L per plot). Additional emergency irrigation was also conducted in June, 2024 with an irrigation rate of about 2 L tree⁻¹ (10 L per plot).

2.4. Tree Row Flora Assessment

In the experiment, living mulch cover on the plot area and the species composition of weeds occurring in the grass sod were analyzed every year, in spring and autumn. A similar, analogous assessment of weed infestation was performed on herbicide fallow plots. Each year it was conducted immediately before herbicide applications in spring, as well as at the end of the vegetation period, in autumn. The only exception was the first assessment, which was made one month after sowing the living mulches, in autumn 2021. The area of a single experimental plot (7.0 m²) in the tree row was narrowed to 5.6 m² to eliminate the edge effect.

To determine the percentage of individual taxa on the plot area, a modified Lipecki and Janisz [42] methodology was used. It was based on a visual assessment of the degree of plot surface coverage by each grass living mulch and identified weed taxa. The scale included the following ranges of percentage soil cover in tree rows: ‘1’—cover in the range of up to 1%, analogously: ‘10’—up to 10%, ‘20’—up to 20%, ‘40’—up to 40%, ‘60’ up to 60%, ‘80’—up to 80%, and ‘100’—up to 100%. When it was possible, recorded weeds were identified at the species level, but in some cases the identification was limited to the genus level. In the case of perennial grasses, some individuals were classified collectively into the botanical family (Poaceae). This group included various grass species, most often they were: Holcus lanatus L., L. perenne L., and Dactylis glomerata L.

Weed taxa identification was based on a phytosociological relevé, which was conducted separately for each experimental plot. It involves a detailed examination of the studied area and preparing a description based on the weed taxa found within it. They were described in terms of the number of identified taxa and the share of soil surface covered by weed taxa in herbicide fallow and living mulch sod. Direct identification was based on the herbology and botany knowledge of two researchers identifying weed flora in the orchard. Additional assistance was provided by a weed taxa database collected over the last 20 years of experiments at the Fruit Experimental Station of the Wrocław University of Environmental and Life Sciences. The taxa list, thus prepared, was used to highlight species present in each assessed experimental plot, and, additionally, newly emerging weeds were added on an ongoing basis. In this way, a dataset was obtained for each plot separately. In addition, weed groups including annual and perennial, as well as mono- and dicotyledonous, plants were assessed.

The nomenclature of the scientific names of vascular plants and their abbreviations (EPPO Code) was based on the taxonomy described by EPPO Global Database [43] which is maintained by the Secretariat of the European and Mediterranean Plant Protection Organization (EPPO). In addition, the plant preferences regarding nitrogen availability were given by Ellenberg et al. [35].

2.5. Statictical Analysis

Statistical analysis of the data was performed using RStudio version 2025.05.0. The data were analyzed by two-way ANOVA for randomized block design from the ‘doebioresearch’ package. Significant differences were marked * at p ≤ 0.05, ** at p ≤ 0.01, *** at p ≤ 0.001. The Shapiro–Wilk test was used to verify if the data followed a normal distribution. For the data describing the share of soil surface under weed taxa cover (expressed as a percentage) that do not fulfill the requirement, a normal distribution was transformed angularly by the function of Bliss or exponential transformations. In cases when a significant ANOVA result was obtained, as well as when it was not obtained, the variables were subjected to further post hoc testing by the Tukey procedure. Significant differences between means were calculated with a Tukey test at p = 0.05.

Principal component analysis (PCA) was performed using the ‘factoextra’ package and visualized using the ‘ggplot2’ package. PCA was applied to evaluate the composition of weed species within the studied functional groups. The analysis was based on data collected from autumn 2021 to autumn 2024 and was performed separately for each year and weed taxa covering >1% and ≤1% of the soil surface. Each subset was analyzed independently, and therefore the PCA axes were scaled separately. The calculations were performed using a correlation matrix, which allowed for visualization of the relationships among weed taxa occurrence patterns.

3. Results

3.1. Living Mulch Sod Cover

In the autumn of the first year of living mulch assessment (2021), none of the grass cover crops used achieved 75% soil surface sod cover (Figure 1). Grass development in the following growing season led to 100% coverage, which was observed until spring 2024. This was influenced by the highest total precipitation in 2023 (Table S1). In the autumn of 2024, a decrease in sod coverage to 70–80% was observed for both FRC treatments.

3.2. Weed Taxa Number and Soil Cover

During the 4 years of the study, a total of 24–31 annual (

Table 1. Total number of identified annual weed taxa in herbicide fallow (HF) and living mulch sod—strong creeping fescue (FRR) or Chewing’s fescue (FRC)—depending on the dose of N (kg ha⁻¹), from autumn 2021 up to autumn 2024.

Weed Taxa		Orchard Floor Management and Dose of Nitrogen (kg ha−1)
		HF		FRR		FRC
Scientific Name	EPPO Code	50	100	50	100	50	100
Amaranthus retroflexus L.	AMARE	+	+	+	+	+	+
Arabidopsis thaliana (L.) Heynh.	ARBTH	+	+	+	–	+	+
Capsella bursa-pastoris (L.) Medik.	CAPBP	+	+	+	+	+	+
Cerastium glomeratum Thuill.	CERGL	+	+	+	–	–	–
Chenopodium album L.	CHEAL	+	+	+	+	+	+
Erigeron canadensis L.	ERICA	+	+	+	+	+	+
Crepis spp.	1CVPG	+	+	+	+	+	+
Digitaria spp.	1DIGG	+	+	+	+	+	+
Draba verna L.	ERPVE	+	+	+	–	–	–
Echinochloa crus-galli (L.) P.Beauv.	ECHCG	+	+	+	+	+	–
Erigeron annuus (L.) Desf.	ERIAN	–	+	–	+	–	+
Euphorbia helioscopia L.	EPHHE	+	–	–	–	–	–
Fumaria officinalis L.	FUMOF	+	+	–	+	+	+
Galinsoga parviflora Cav.	GASPA	+	+	+	+	+	+
Galium aparine L.	GALAP	–	–	+	–	–	–
Geranium spp.	1GERG	+	+	+	+	+	+
Lamium amplexicaule L.	LAMAM	+	+	+	+	+	+
Lamium purpureum L.	LAMPU	+	+	+	+	+	+
Lactuca serriola L.	LACSE	–	+	+	+	+	+
Matricaria chamomilla L.	MATCH	+	+	+	+	+	+
Matricaria discoidea DC.	MATMT	–	–	–	–	+	–
Medicago lupulina L.	MEDLU	+	+	+	+	+	+
Poa annua L.	POAAN	+	+	+	+	+	+
Polygonum aviculare L.	POLAV	+	–	–	–	+	+
Senecio vulgaris L.	SENVU	+	+	+	+	+	+
Setaria spp.	1SETG	+	+	+	+	–	+
Silybum marianum (L.) Gaertn.	SLYMA	+	+	–	–	–	–
Sonchus asper (L.) Hill.	SONAS	+	+	+	+	–	+
Sonchus oleraceus L.	SONOL	+	+	+	+	+	+
Stellaria media (L.) Vill.	STEME	+	+	+	+	+	+
Sisymbrium officinale (L.) Scop.	SSYOF	–	+	–	–	–	–
Urtica urens L.	URTUR	–	+	–	–	–	–
Veronica arvensis L.	VERAR	+	+	+	+	+	–
Veronica persica Poir.	VERPE	+	+	+	+	–	+
Veronica sp.	1VERG	–	–	–	+	+	–
Vicia spp.	1VICG	–	+	–	–	+	–
Viola arvensis Murr.	VIOAR	+	–	+	+	+	–
Total number of weed taxa		29	31	27	26	26	24

) and 11–20 perennial (

Table 2. Total number of identified perennial weed taxa in herbicide fallow (HF) and living mulch sod—strong creeping fescue (FRR) or Chewing’s fescue (FRC)—depending on the dose of N (kg ha⁻¹), from autumn 2021 up to autumn 2024.

Weed Taxa		Orchard Floor Management and Dose of Nitrogen (kg ha−1)
		HF		FRR		FRC
Scientific Name	EPPO Code	50	100	50	100	50	100
Achillea millefolium L.	ACHMI	+	+	+	+	+	+
Artemisia vulgaris L.	ARTVU	–	–	+	–	–	–
Bellis perennis L.	BELPE	+	+	+	–	+	–
Calamagrostis sp.	1CLMG	+	+	+	+	+	+
Cerastium holosteoides Fr.	CERVU	+	+	+	+	+	+
Cirsium arvense (L.) Scop.	CIRAR	+	+	+	+	+	+
Cirsium vulgare (Savi) Ten.	CIRVU	+	+	–	–	+	–
Convolvulus arvensis L.	CONAR	+	+	+	+	+	+
Elymus repens (L.) Gould	AGRRE	+	+	+	+	+	+
Epilobium spp.	1EPIG	+	+	–	–	–	–
Hieracium spp.	1HIEG	–	–	+	–	–	–
Malva neglecta Wallr.	MALNE	–	+	–	–	–	–
Malva sylvestris L.	MALSI	+	+	+	+	+	+
Potentilla argentea L.	PTLAG	+	+	–	–	–	–
Plantago major L.	PLAMA	–	–	+	+	–	–
Poaceae—other species	1GRAF	+	+	+	+	+	+
Rumex spp.	1RUMG	–	–	–	–	–	+
Solidago sp.	1SOOG	+	+	–	–	–	–
Tanacetum vulgare L.	CHYVU	–	+	–	–	+	–
Taraxacum spp.	1TARG	+	+	+	+	+	+
Trifolium hybridum L.	TRFHY	–	+	–	–	–	–
Trifolium medium L.	TRFME	–	–	–	–	+	–
Trifolium ochroleucon Huds.	TRFOC	–	–	–	–	–	+
Trifolium pratense L.	TRFPR	–	+	+	–	+	+
Trifolium repens L.	TRFRE	+	+	+	+	+	+
Urtica dioica L.	URTDI	–	+	–	–	–	–
Total number of weed taxa		15	20	15	11	15	13

) weed taxa were identified depending on the studied treatment. The highest mean number of weed taxa was recorded in both HF treatments in each year of flora assessment (Figure 2). In some cases, it reached approximately 20. In living mulch treatments, the number of weed taxa was dependent on the cover of grass subspecies. From the beginning of the third year of the study (2023), fewer taxa appeared in the FRR, regardless of the nitrogen rate. Good sod development in the spring of 2023 suppressed the number of taxa, sometimes to fewer than four. The species diversity of weed taxa in the year of sowing the mulches (2021) was low among all treatments (Table S2). However, there was a slight decrease in the diversity of taxa in the two living mulches, which was observed more clearly in spring 2022 (Table S3). The mean number of weed taxa covering more than 1% of the soil surface was significantly higher in HF than in FRR and FRC sod (

Table 3. Characteristics of weed taxa populations occurring in herbicide fallow (HF) and living mulch sod—strong creeping fescue (FRR) or Chewing’s fescue (FRC)—depending on the dose of N (kg ha⁻¹), 2021–2024 mean.

Orchard Floor Management and Dose of Nitrogen (kg ha−1)		Mean Number of Taxa	Mean Number of Taxa Which Cover		Share of Soil Surface Under Weed Taxa Cover (%)
			≤1%	>1%	Annual	Perennial	Monoco-Tyledons	Dico-Tyledons
			of Soil Surface
Means for treatments
HF	50	16.2 b	7.4 a	8.7 b	79.3 b	41.5 c	57.5 b	77.5 b
	100	16.2 b	7.3 a	8.9 b	85.7 b	34.0 bc	50.8 b	80.7 b
FRR	50	8.8 a	6.0 a	2.8 a	19.4 a	16.3 a	9.0 a	29.3 a
	100	9.4 a	6.6 a	2.8 a	20.1 a	14.3 a	13.1 a	26.6 a
FRC	50	10.1 a	6.9 a	3.2 a	18.5 a	25.2 ab	11.9 a	33.3 a
	100	9.6 a	6.2 a	3.4 a	21.9 a	20.4 a	9.6 a	35.1 a
Statistical significance D × OFM		NS	NS	NS	NS	NS	NS	NS
Means for nitrogen doses
50		11.7 a	6.8 a	4.9 a	39.0 a	27.7 b	26.1 a	46.7 a
100		11.7 a	6.7 a	5.0 a	42.6 a	22.9 a	24.5 a	47.5 a
Statistical significance		NS	NS	NS	NS	*	NS	NS
Means for orchard floor managements
HF		16.2 b	7.3 a	8.8 b	82.5 b	37.7 c	54.1 b	79.1 b
FRR		9.1 a	6.3 a	2.8 a	19.7 a	15.3 a	11.0 a	28.0 a
FRC		9.8 a	6.5 a	3.3 a	20.2 a	22.8 b	10.7 a	34.2 a
Statistical significance		***	NS	***	***	***	***	***

*, ***—significantly different (ANOVA) at p ≤ 0.05 and p ≤ 0.001, respectively; NS—not significantly different. Means marked with different letters in separate columns represent statistical differences among eight treatments (D × OFM) or two N doses (D) or three floor managements (OFM) (Tukey test, p = 0.05).

). The proportions of sporadic weeds that covered 1% or less of the soil surface were not statistically different. There was no significant effect of nitrogen dose on the total number of weed taxa throughout the study period. The mean number of weed taxa (2021–2024) was the same—11.7 for the doses of 50 and 100 kg N ha⁻¹ (Table 3). However, a significant increase in the number of taxa was observed in HF compared to living mulches. Nitrogen dose had no significant effect on weed population characteristics, except for the share of soil surface under perennial weed cover. On the contrary, orchard floor management significantly influenced the mean number of taxa and the share of soil surface cover under annual and perennial as well as mono- and dicotyledonous plants. Significantly higher populations were observed in the HF compared to the FRR and FRC plots.

Analysis of weed populations in subsequent years showed that the percentage of soil surface under annual weed taxa cover was high in HF (Figure S1). In some seasons, it reached 100%. However, the use of living mulch effectively reduced this population from autumn 2022. For perennial weed taxa, higher cover was also noted in HF (Figure S2). Maximum cover reached 65% of the soil surface. In living mulch sod, perennial weeds did not exceed 33%, with the exception of FRC sod with a dose of 50 kg N ha⁻¹ (60%). Characteristics of weed taxa populations in terms of the occurrence of mono- and dicotyledonous plants showed a large diversity of the studied orchard floor management (Figure S3). The initial coverage of the soil surface with monocotyledonous populations in HF was just over 10% (autumn 2021). In the following years of study, an increase was noted, reaching up to 90% (autumn 2022). In contrast, in the FRR and FRC sod, the coverage did not exceed 40%. In some seasons, their share was even marginal. Dicotyledonous weed taxa also dominated in HF (Figure S4). During the first assessment in 2021, soil cover was approximately 25%. In subsequent years, the coverage intensity reached up to 100% (autumn 2024). In the grass living mulch, dicotyledonous weed taxa dominated in spring 2022. In the FRC sod, their coverage was slightly greater compared with the FRR sod. For both mono- and dicotyledonous populations, no effect of nitrogen dose on the percentage of soil surface under them was observed.

3.3. PCA Results

PCA showed a high dispersion of weed taxa covering more than 1% of the soil surface between soil treatments in tree rows in autumn 2021 (Figure 3). Among sporadically occurring weeds (≤1%), mainly several annual weed taxa as well as Convolvulus arvensis L. were associated with HF. Already in 2022, a very clear division was visible in terms of the occurrence of taxa covering more than 1% of the area (Figure 4, Figure 5 and Figure 6). The four studied combinations with grass living mulch formed a separate group associated with a small number of taxa compared to their richness in HF 50 and HF 100. In 2022, living mulch treatments showed strong associations with taxa such as C. arvensis L., Achillea millefolum L., and both annual Lamium species (Figure 4). This division was also noted in 2023 (Figure 5), when grass living mulch treatments were associated with perennial weed taxa occurring the previous year, joined by Trifolium repens L. and Cirsium arvense L. In 2024, however, the weeds that had previously dominated the grass sod dispersed (Figure 6). Only A. millefolium L. and Calamagrostis sp. remained associated with the use of living mulch in the tree rows. PCA showed a trend toward differentiation of the weed community between HF 50 and HF 100. Over the years (2022–2024), taxa covering 1% or less of the soil surface showed no clear association with orchard floor management. Weed species compositions were diverse across all orchard floor management treatments, with no clear association with nitrogen fertilization.

3.4. Dominant Weed Taxa

T. repens L., over the four years of study (2021–2024), was characterized by a significantly higher average percentage of soil surface cover in HF compared to grass living mulches (

Table 4. Mean percentage of soil surface under the most important perennial weed taxa cover occurring in herbicide fallow (HF) and living mulch sod—strong creeping fescue (FRR) or Chewing’s fescue (FRC)—depending on the dose of N (kg ha⁻¹), 2021–2024 mean.

Orchard Floor Management and Dose of Nitrogen (kg ha−1)		Weed Taxa
		ACHMI	1CLMG	CONAR	MALSI	1GRAF	1TARG	TRFRE
Means for treatments
HF	50	0.1 a	0.4 a	0.1 a	7.8 a	6.4 a	16.0 c	26.9 b
	100	0.3 a	0.4 a	1.9 a	11.0 a	11.3 a	9.1 bc	14.4 a
FRR	50	2.0 a	2.4 a	4.0 a	0.3 a	0.9 a	2.9 ab	9.2 a
	100	0.2 a	3.4 a	0.1 a	3.5 a	1.6 a	0.9 a	5.4 a
FRC	50	6.8 a	6.7 a	5.6 a	3.4 a	1.6 a	1.6 a	7.5 a
	100	0.8 a	1.9 a	0.4 a	10.6 a	1.2 a	2.2 a	6.5 a
Statistical significance D × OFM		NS	NS	NS	NS	NS	NS	*
Means for nitrogen doses
50		2.9 b	3.1 a	3.2 a	3.8 a	2.9 a	6.8 a	14.5 b
100		0.4 a	1.9 a	0.8 a	8.4 a	4.7 a	4.1 a	8.8 a
Statistical significance		*	NS	NS	NS	NS	NS	**
Means for orchard floor managements
HF		0.2 a	0.4 a	1.0 a	9.4 a	8.8 b	12.6 b	20.7 b
FRR		1.1 a	2.9 ab	2.0 a	1.9 a	1.3 a	1.9 a	7.3 a
FRC		3.8 a	4.3 b	3.0 a	7.0 a	1.4 a	1.9 a	7.0 a
Statistical significance		NS	*	NS	NS	***	***	***

*, **, ***—significantly different (ANOVA) at p ≤ 0.05, p ≤ 0.01, and p ≤ 0.001, respectively; NS—not significantly different. Means marked with different letters in separate columns represent statistical differences among eight treatments (D × OFM) or two N doses (D) or three floor managements (OFM) (Tukey test, p = 0.05).

). The spread of this species was significantly influenced by orchard floor management, nitrogen dose, and the interaction of both factors. Similarly, Taraxacum spp. was a significant perennial weed in HF. However, the nitrogen dose did not have a significant effect on the share of the soil surface cover. This was similar for other perennial weeds, with the exception of A. millefolium L. A significant effect of the second factor studied—floor management—was observed in the case of Taraxacum spp. and monocotyledonous taxa represented by Calamagrostis sp. and a group of perennial species from the Poaceae family. Populations of these weeds were significantly lower in grass living mulch.

All identified annual species, except Geranium spp., had a significantly higher mean percentage of soil surface in HF compared to grass living mulches (

Table 5. Mean percentage of soil surface under the most important annual weed taxa cover occurring in herbicide fallow (HF) and living mulch sod—strong creeping fescue (FRR) or Chewing’s fescue (FRC)—depending on the dose of N (kg ha⁻¹), 2021–2024 mean.

Orchard Floor Management and Dose of Nitrogen (kg ha−1)		WEED TAXA
		CAPBP	CHEAL	1DIGG	1GERG	LAMPU	STEME	POAAN
Means for treatments
HF	50	25.2 b	5.7 a	3.8 a	7.5 a	6.5 a	55.7 b	55.0 c
	100	22.7 b	7.5 a	5.1 a	4.0 a	3.8 a	66.8 b	43.6 b
FRR	50	1.3 a	1.1 a	0.9 a	1.4 a	1.2 a	10.2 a	6.1 a
	100	0.7 a	1.8 a	1.9 a	2.3 a	1.6 a	12.9 a	7.5 a
FRC	50	0.6 a	1.9 a	1.2 a	2.8 a	0.9 a	10.3 a	3.3 a
	100	1.7 a	1.9 a	1.5 a	1.4 a	2.6 a	15.7 a	5.7 a
Statistical significance D × OFM		NS	NS	NS	NS	NS	NS	*
Means for nitrogen doses
50		9.0 a	2.9 a	2.0 a	3.9 a	2.9 a	25.4 a	21.5 a
100		8.3 a	3.8 a	2.8 a	2.6 a	2.7 a	31.8 b	18.9 a
Statistical significance		NS	NS	NS	NS	NS	*	NS
Means for orchard floor managements
HF		23.9 b	6.6 b	4.5 b	5.8 a	5.2 b	61.3 b	49.3 b
FRR		1.0 a	1.5 a	1.4 a	1.8 a	1.4 a	11.6 a	6.8 a
FRC		1.1 a	2.0 a	1.3 a	2.1 a	1.8 a	13.0 a	4.5 a
Statistical significance		***	***	*	NS	**	***	***

*, **, ***—significantly different (ANOVA) at p ≤ 0.05, p ≤ 0.01, and p ≤ 0.001, respectively; NS—not significantly different. Means marked with different letters in separate columns represent statistical differences among eight treatments (D × OFM) or two N doses (D) or three floor managements (OFM) (Tukey test, p = 0.05).

). A significant effect of nitrogen dose was noted only in the case of the species Stellaria media (L.) Vill., which had a greater coverage on the plots with a dose of 100 kg N per ha⁻¹.

4. Discussion

4.1. F. rubra L. as Living Mulch

F. rubra L., along with several other grass species such as L. perenne L., P. pratensis L., or Festuca ovina L., is considered one of the basic grasses used for covering tractor alleys in an orchard [4]. Using this species as living mulch extended its cultivation area by including tree rows, which ensured complete soil coverage in the orchard. Two subspecies of F. rubra L., with extravaginal branching of tillers [44], demonstrated strong weed suppression in our experiment. FRR, which has a creeping growth habit, develops well-formed rhizomes [15] and maintained full (100%) soil cover in the tree rows by the end of our study. However, this subspecies was classified at an intermediate level in terms of its suitability as a lawn grass [19]. The criteria included turf density, slow re-growth, and winter hardiness. In another assessment of sod cover, FRR received a lower value compared to FRC [18]. In contrast, in our study, a decline in soil cover was observed only in FRC, appearing three years after sowing (autumn 2024). This suggests a gradual decline of this living mulch in the tree rows. Nevertheless, both subspecies of F. rubra L. performed well as cover crops in the young apple orchard. The space necessary for the growth and development of annual species [45] was filled by living mulch, consistent with results of other studies of the same species [21,22]. The sod significantly prevented the development of troublesome perennial weeds such as Taraxacum spp. and T. repens L. and also limited the spread of other species from the Poaceae family.

4.2. Nitrogen Fertilization in Tree Rows with Living Mulch

In accordance with the current Methodology of Integrated Apple Production [39], in our experiment we adopted an individual broadcast method of applying nitrogen directly under the tree canopies. This reduced losses of nitrogen across the entire orchard area. First of all, it limited the area of grass sod fertilized with nitrogen. Assuming that ammonium nitrate granules were distributed within a radius of approximately 30 cm around the tree canopy, the actual area of sod subject to fertilization was approximately 20% of the whole area in the tree rows. This helps to explain the lack of a significant effect of doubling the nitrogen dose from 50 to 100 kg on the percentage of soil coverage of the most important annual species. Among the annual taxa, the only exception was S. media (L.) Vill.—a species that acts as a nitrogen indicator plant [35]. It reached the highest average soil cover among annual taxa weeds and its presence directly under the young tree canopies in the fertilized zone was therefore likely.

Among perennial weeds, no significant effect of nitrogen on the most important taxa was found, even though species such as Taraxacum spp., from the point of view of Ellenberg et al. [35], belong to nitrophilous plants. The assessment of the whole perennial population showed a greater spread at the lower applied dose—50 compared with 100 kg N ha⁻¹. This may suggest that their presence in the tree rows was not directly related to the area of N fertilization. The absence of a clear effect of doubling the nitrogen dose from 50 to 100 kg ha⁻¹ on the growth of perennial weed populations in grass living mulch may be advantageous for fruit-growing practice, as it indicates the possibility of limiting competition from living mulch and the weed taxa present within it. However, the problem of nitrogen uptake by F. rubra L. itself remains unresolved, as it absorbs nitrogen despite being relatively indifferent to soil nitrogen availability [35]. To avoid such nitrogen losses, it may be necessary to introduce only foliar fertilization to trees in orchards with living mulch.

4.3. The Number of Weed Taxa as a Measure of Biodiversity in Tree Rows

Biodiversity in commercial orchards may range from 18 to 28 weed taxa [46]. As expected, a rich weed composition was observed in the herbicide fallow, similar to other studies [30,31], or in a mixed system with chemical and mechanical weed control [25]. In our study, the total number of identified weed taxa in herbicide fallow ranged from 44 (HF 50) to 51 (HF 100), indicating greater biodiversity in this orchard floor management method compared to both grass living mulches. In an average year of the study, directly before herbicide applications, the mean number of taxa under the apple tree canopy was about 16. Most of these taxa occurred sporadically and covered no more than 1% of the soil surface in the tree rows in the year of orchard establishment and in the spring of 2022. Compared to herbicide fallow, the weed composition in grass living mulch was less diverse, but remained constant over time, due to the absence of chemical interference in the grass sod. Throughout the study period, a similar number of annual and perennial taxa was identified in the sod of both FRR and FRC, ranging from 37 (FRR 100) to 42 (FRR 50). This means that the mean number of taxa was significantly lower in the grass cover crop (fewer than 10) compared to the herbicide fallow (about 16). In other studies, the number of recorded species within different cover crops ranged from 25 to 31 [12], but in a single year of study, 12 to 16 weed taxa appeared in a single-species living mulch [47]. In several commercial orchards, the number of taxa in different cover crops ranged from 4 to 21 [9]. In 2022 and 2023, when living mulches filled the entire tree row area, in our experiment, the share of annual species associated with this floor management method was marginal. The perennial population was also suppressed by the presence of grass cover crops. The first few years of F. rubra L. presence as living mulch limited weed infestation, but this period of grass cultivation did not favor increased biodiversity in the apple tree rows.

4.4. Interaction Between Strong Creeping and Chewing’s Fescue with Dose of Nitrogen vs. Weed Community

Although nitrogen fertilization and the use of F. rubra L. subspecies as living mulch have both been studied individually in orchard management, little research has directly examined their combined effect on shaping weed communities. The total numbers of identified annual and perennial taxa recorded in FRR and FRC sod were very similar. However, during the most dynamic development phase of both living mulches (between spring 2023 and spring 2024), when each sod covered 100% of the soil surface, the mean number of taxa was lower in FRR compared to FRC, likely due to the differences in perennial taxa. This trend was supported by the share of soil surface covered by perennial weeds, which was significantly lower in FRR than in FRC. However, the interaction of the two grass species with two nitrogen doses did not result in a significant difference in perennial or annual weed cover.

It is worth noting once again that nitrogen fertilization was applied to support apple trees and was limited to the soil directly beneath the tree canopies. Its effect therefore covered only a small part of the tree rows, which may explain the limited effect of the interaction of this factor with strong creeping and Chewing’s red fescue sod on weed taxa community and its cover in our experiment. Research by DeBels et al. [37] on Chewing’s fine fescue showed that only the application of a dose of 196 kg N ha⁻¹ allowed for a significant reduction in the percentage of weed cover compared to the unfertilized control. In contrast, a case study of one weed species—P. annua L.—by Braun et al. [36] found no significant interaction between three subspecies of F. rubra L. and several doses of nitrogen fertilization on the percentage cover of this weed. Furthermore, as described in the publications cited above, the combination of F. rubra L. subspecies with nitrogen fertilization has been conducted in the context of turfgrass management and not in the fruit-growing context, where ecological interactions and management goals differ substantially. This lack of evidence highlights the need for further research to clarify how these two experimental factors jointly influence weed composition and competition dynamics in perennial cropping systems, such as orchards.

5. Conclusions

In the period stricte between herbicide applications, numerous annual taxa appeared in herbicide fallow and perennial taxa tolerant of the agrochemicals used grew back. The use of both subspecies of F. rubra L. limited the number of taxa in tree rows, especially those occurring sporadically, with coverage not exceeding 1% of the soil surface. This was related to the dynamic development of F. rubra L subspecies sod, which effectively reduces the space for germination and further growth of annual weed taxa, as well as limited the spread of perennial plants, compared with herbicide fallow. PCA showed a clear differentiation in species composition between herbicide fallow and living mulch and also indicated a tendency towards greater stability of the grass living mulch agroecosystem. This was manifested by lower variability of weed taxa in sod compared to the abundant weed composition in the herbicide fallow. The application of two doses of nitrogen, 50 and 100 kg ha⁻¹, stricte under the canopies of trees, most often had no significant effect on the number of weed taxa and their soil surface coverage in the rows of apple trees. Importantly, doubling the nitrogen fertilization rate, while limiting the application area to the tree canopy, did not increase the perennial weed population in the living mulch sod. This method can be recommended for orchards, with the aim of increasing the nitrogen supply to fruit trees.

A comparison of the two subspecies, strong creeping red fescue and Chewing’s red fescue, from the perspective of reducing weed taxa cover under tree canopies allows us to assess their suitability as living mulches in apple tree rows. However, their use does not contribute to enriching the biodiversity of the flora of the orchard agroecosystem.

The significantly higher average share of soil surface covered by perennial weed taxa in Chewing’s red fescue, together with its sod cover decrease, already in the fourth year of cultivation, compared to strong creeping red fescue, is unfavorable for its adaptation as living mulch. This suggests that the bunch-type red fescue subspecies may have limited usefulness in subsequent years of orchard management.