Evaluating soil cover strategies for enhancing water conservation, biomass contribution, and weed control in rocoto pepper (Capsicum pubescens Ruiz & Pav.) cultivation under arid conditions

Poma-Chamana, Russell; Cama-Moreno, Elick; Flores-Marquez, Ricardo; Quello, Antony; Solórzano, Richard

doi:10.3389/fagro.2025.1663633

ORIGINAL RESEARCH article

Front. Agron., 11 November 2025

Sec. Field Water Management

Volume 7 - 2025 | https://doi.org/10.3389/fagro.2025.1663633

This article is part of the Research TopicEnhancing Agricultural Water Management: Techniques for Improving Crop Water Efficiency and Sustainability, Volume IIView all 5 articles

Evaluating soil cover strategies for enhancing water conservation, biomass contribution, and weed control in rocoto pepper (Capsicum pubescens Ruiz & Pav.) cultivation under arid conditions

Russell Poma-Chamana¹

Elick Cama-Moreno¹

Ricardo Flores-Marquez²

Antony Quello¹

Richard Solórzano^2,3*

¹Dirección de Servicios Estratégicos Agrarios, Estación Experimental Agraria de Arequipa, Instituto Nacional de Innovación Agraria (INIA), Arequipa, Peru
²Dirección de Servicios Estratégicos Agrarios, Instituto Nacional de Innovación Agraria (INIA), Lima, Peru
³Facultad de Ciencias Ambientales, Universidad Científica del Sur (UCSUR), Lima, Peru

Sustainable crop production in arid regions is challenged by soil moisture loss, low organic matter content, and intense weed competition. This study aimed to assess the impact of different mulching strategies on the main crop yield, soil moisture conservation, the contribution of dry biomass and nitrogen to the soil by living covers, and the efficacy of weed control. The experiment was conducted in Santa Rita de Siguas (Arequipa, Peru) using a completely randomized block design with four treatments and three replicates. Measurements included biometric and physiological parameters of rocoto pepper (Capsicum pubescens), volumetric soil moisture, dry biomass, and nitrogen content in living covers, as well as weed density. Results indicated that most treatments had no significant impact on the biometric and physiological parameters of rocoto. Plastic mulch reduced irrigation demand and suppressed weeds, although crop yield did not significantly differ among treatments. Among the living mulches, Trifolium pratense was more effective than Melilotus albus in conserving soil moisture, maintaining levels up to 15.86%. In contrast, Melilotus albus produced the highest above-ground dry biomass (8.57 t·h⁻¹), although both legume species accumulated similar amounts of nitrogen in their biomass. Both living covers gradually reduced weed populations, though without complete eradication. In conclusion, plastic mulch represents a potential option under conditions of severe water limitation. Meanwhile, leguminous cover crops, particularly M. albus, offer an alternative complementary strategy for enhancing soil organic matter and could promote long-term sustainability of the cropping system. These findings warrant extended temporal validation to confirm their reproducibility and reliability.

1 Introduction

The rocoto pepper (Capsicum pubescens Ruiz & Pav.) is one of the five domesticated species within the Capsicum genus. It is native to the arid regions of the Peruvian-Bolivian Andes and later expanded into tropical regions (Walsh and Hoot, 2001). Peru holds the greatest genetic diversity of this species, with 712 preserved accessions, 299 of which belong to C. pubescens. Many of these are conserved at the Arequipa Agrarian Experimental Station (Libreros et al., 2013). These accessions are classified into two primary genotypes, serrano and monte, which are distinguished by differences in fruit size and pungency (Hernández-Amasifuen et al., 2022). Among the monte accessions, the red rocoto is particularly notable for its commercial significance in Peru, with an annual production of up to 54,000 tons, 80% of which is concentrated in the central jungle, especially in the Pasco department (MIDAGRI, 2022; Lozano Alarcón, 2012).

In the Arequipa region, the consumption of red rocoto is particularly significant, as it serves as the main ingredient in the traditional dish ‘rocoto relleno’ and various other culinary preparations (Fuentes and del, 2014; Blanco de Alvarado-Ortiz, 2016). However, the monte red rocoto is not commercially cultivated in this region, despite similar temperature ranges between the province of Oxapampa, the main rocoto-producing area, and the province of Arequipa. Historical average temperatures range from a maximum of 24.78°C and a minimum of 13.55°C in Oxapampa to a maximum of 32.2°C and a minimum of 9°C in Santa Rita de Siguas, Arequipa (SENAMHI - Descarga de Datos).

Agriculture in arid regions is subject to severe water limitations. In Arequipa, 25,942 hectares out of 146,824 hectares of agricultural land are located in arid zones, including Majes (15,932 ha), Santa Rita de Siguas (3,032 ha), and La Joya (6,978 ha), all of which rely exclusively on irrigation (Superficie Agrícola). This dependence results in critical water constraints, particularly in the Majes irrigation system, where high water-demand crops contribute to recurrent water deficits (Zapana Churata, 2018). The situation is further exacerbated by inefficient water management, which leads to low irrigation efficiency and substantial water losses through deep infiltration (Wei et al., 2021). Additionally, these arid areas present low organic matter content and reduced water retention capacity (Weil and Brady, 2017) However, proper irrigation management has the potential to increase soil organic carbon (Trost et al., 2013) and improve the cation exchange capacity of coarse-textured soils (Weil and Brady, 2017). These challenges highlight the need to evaluate agronomic strategies, such as mulching use, to optimize water management and organic carbon sequestration in rocoto pepper production in Arequipa.

Mulching involves covering the soil surface with organic or inorganic materials to improve soil properties and optimize crop development (Mulching (Ac.) (484) Conservation Practice Standard | Natural Resources Conservation Service). This practice increases soil organic matter, reduces bulk density, enhances aggregate stability, increases water retention, and regulates soil temperature (Mulching (Ac.) (484) Conservation Practice Standard | Natural Resources Conservation Service; Alyokhin et al., 2019). Additionally, mulching can contribute to pest control and increase crop yields (Cai et al., 2022; Muhammad et al., 2022). Plastic mulch is particularly effective in reducing soil evaporation, increasing soil temperature, and providing efficient weed control (Sokombela et al., 2025), whereas living mulches promote soil biological activity and enhance nutrient availability through the decomposition of plant biomass (Kołota and Adamczewska-Sowińska, 2013). Leguminous cover crops are particularly valuable for their ability to fix atmospheric nitrogen, enrich the soil with high-quality organic matter, and improve nutrient cycling and water retention (Stagnari et al., 2017).

Trifolium pratense L. and Melilotus albus Medik exhibit high agronomic potential. T. pratense is a perennial species native to temperate regions, characterized by variable root architecture. Erect plants develop deep main roots, whereas low-growing forms produce branched root systems with secondary roots near the stem base (Taylor and Quesenberry, 1996), thereby optimizing water uptake based on the distribution of soil moisture. Commercial cultivars typically produce greater biomass than wild types (Heslop et al., 2025) and have a high nitrogen fixation capacity, reaching up to 545 kg N·ha⁻¹·year⁻¹, along with notable weed control potential (Guntli et al., 1999; Anglade et al., 2015). M. albus is an annual or biennial species with erect stems ranging from 30 to 180 cm, capable of producing substantial above-ground biomass (12 t·ha⁻¹ of dry matter), making it a promising candidate for restoring degraded soils (Țîței, 2022). This species is well-adapted to saline grasslands and arid environments, although its ecology remains largely unexplored (Zabala et al., 2018; Wang et al., 2024), and it has demonstrated adaptability to the desert conditions of Arequipa (Quipuscoa Silvestre et al., 2016).

The application of mulching in rocoto (Capsicum pubescens) cultivation remains poorly researched. In related Capsicum species, live mulching with Trifolium pratense has been shown to provide superior soil moisture retention under water-stress conditions compared to conventional tillage (Biazzo and Masiunas, 2000). Meanwhile, plastic mulch has significantly increased yields under arid conditions (Odokonyero et al., 2025). In this regard, no studies have directly compared the effects of plastic mulching and leguminous living mulches on this crop in Peru, despite the economic importance of rocoto peppers and the cultivation challenges posed by arid regions. This gap proves particularly significant given the dual need to optimize water use while enhancing soil carbon sequestration in these environments. Within this context, the present study aims to evaluate the impact of different mulching strategies (plastic mulching versus leguminous living mulches) on crop yield, soil moisture conservation, dry biomass and nitrogen inputs to soil, as well as weed control in rocoto pepper cultivation under arid conditions.

2 Materials and methods

2.1 Study area

The study was conducted at the Arequipa Agrarian Experimental Station of the National Institute for Agrarian Innovation (INIA), located in the district of Santa Rita de Siguas, province and department of Arequipa, Peru. The experimental plot was situated at 16°28′22.0″ S, 72°06′35.9″ W, at 1,274.28 m.a.s.l. (Figure 1). The region experiences an average annual rainfall of 0.11 mm, with minimum temperatures of 9.3°C in July and maximum temperatures ranging from 25 to 26°C between September and April. These values were calculated based on historical records (1949–2013) of the Majes weather station (16°20′8.4″ S, 72°9′8.92″ W), managed by the National Service of Meteorology and Hydrology of Peru (SENAMHI) (SENAMHI - Descarga de Datos). Additionally, meteorological data were obtained during the experimental period from the Santa Rita weather station, also operated by SENAMHI (Figure 2).

Figure 1

Map and photos of an experimental plot in Santa Rita de Siguas District, Arequipa, Peru. The left panel shows a satellite map with a red circle for the experimental plot and a yellow diamond for the weather station. Inset maps illustrate Peru and Arequipa Province. Panel A displays farmland with a sign describing the experimental project. Panel B presents a diagram of treatment plots labeled T1 to T4 in three blocks. A legend explains colors and symbols used.

Figure 1. Study area. Location of the experimental plot and weather station at the Arequipa Agrarian Experimental Station – INIA, Santa Rita de Siguas district, Arequipa province, Arequipa region, Peru. (A) Image of the surrounding area of the experimental plot of the rocoto (Capsicum pubescens Ruiz & Pav.) plot during the fruiting stage. (B) Layout of the experimental plot.

Figure 2

Graph showing rainfall, ETo, and temperature data from May 2024 to April 2025. Maximum and minimum temperatures are indicated with dashed and dotted lines, respectively. ETo is shaded in green, precipitation in grey, and arrows marked T, BF, FS, BH, and LH point to specific phenological stage.

Figure 2. Meteorological parameters during the 2024–2025 growing season. Potential evapotranspiration (ET₀), precipitation, maximum temperature (Tmax), and minimum temperature (Tmin) during crop development. Blue arrows indicate the beginning of each phenological stage: transplanting (T); beginning of flowering (BF); fruiting stage (FS); beginning of harvest (BH); and last harvest (LH).

2.2 Experimental design

The experiment was arranged in a randomized complete block design (RCBD) with four treatments and three replicates, resulting in a total of 12 experimental units, each measuring 6 × 5 m, with 1m-wide alleys separating the blocks. The planting layout, measuring 1.2 × 0.8 m, provided a planting density of 10,416 plants·ha⁻¹. The treatments were as follows: T1 – soil cover with Melilotus albus; T2 – soil cover with Trifolium pratense; T3 – plastic mulch; and T4 – manual weeding (control). A RCBD was employed to control for potential longitudinal edaphic variability across the experimental field.

2.3 Physicochemical characteristics of the soils

Prior to the establishment of the experimental plot, a composite soil sample was collected from the topsoil layer (0–30 cm) and analyzed at the Soil, Water, and Foliar Analysis Laboratory (LABSAF) of INIA Arequipa to determine its physicochemical properties. The soil was characterized as loamy sand, comprising 80.9% sand, 11.3% silt, and 7.9% clay, as determined by the Bouyoucos hydrometer method (Bouyoucos, 1962). The pH was 7.4 (ISO, 1982), and electrical conductivity (EC 1:5) was 1.05 dS·m⁻¹ (Standardization, 1994). Organic matter content was 1.30% (Walkley and Black, 1934), available phosphorus (P) was 30.40 mg·kg⁻¹ (Olsen and Sommers, 1982), and available potassium (K) was 598.80 mg·kg⁻¹ (DOF - Diario Oficial de la Federación).

2.4 Experimental plot management

The experiment was conducted between May 2024 and March 2025. The land was prepared by subsoiling, ploughing, and leveling using a John Deere 6110D tractor, forming raised beds 0.30 m in heigh and spaced 1.20 m from center to center, resulting in an effective bed width of 0.80 m. A drip irrigation system was installed, consisting of a 2″ main line, 1″ distribution lines, and 16 mm lateral lines with 1.5 L·h⁻¹ emitters at 60 kPa every 0.20 m, along with a fertigation arc equipped with ¾″ Wade Rain Venturi injectors.

Prior to transplanting, the beds were irrigated for one hour. Forty-day-old rocoto seedlings with 4–6 true leaves of the accession ‘rocoto rojo de la selva central’, obtained from a local nursery, were used. Plants were transplanted in a single row at the center of each bed, maintaining a spacing of 0.80 m between plants, which extrapolated to a density of 10,416 plants·ha⁻¹.

Based on the soil characteristics, a fertigation dose of 200–100–200 kg·ha⁻¹ of N (as ammonium nitrate), P₂O₅ (as monoammonium phosphate), and K₂O (as potassium nitrate) was applied at seven-day intervals, starting two days after transplanting (DAT). The fertilization was split according to the requirements of each phenological stage: vegetative growth (40%-60%-15% of N-P₂O₅-K₂O), flowering (25%-20%-30%), fruiting (20%-15%-30%), and maturation (15%-5%-25%).

At 120 DAT, staking was performed by placing two 1.2-meter-high posts at the ends of each bed, connected with ropes. Additionally, an individual 1.2-meter stake was installed for each plant. These practices aimed to ensure plant stability and minimize the risk of fruit-soil contact.

Pest and disease control of the crop was carried out based on field evaluations. For the management of Thrips spp., Bemisia tabaci, and Myzus persicae, LOVERA^® (lambda-cyhalothrin + thiamethoxam; 20 mL per 20 L backpack sprayer; applied at 26, 64, 105, 148, 188, 225 and 243 DAT) and KRAKEN^® (imidacloprid + lambda-cyhalothrin; 20 g per 20 L; applied at 47, 84, 125, 170, 214 and 238 DAT) were sprayed alternately. CIPERMEX^® Super 10 CE (alpha-cypermethrin; 30 mL per 20 L; applied at 63 and 98 DAT) was used for larval control. To manage Tetranychus urticae, DK-TINA^® (abamectin; 15 mL per 20 L) was applied to both the main crop and Trifolium pratense living cover at 148, 217, 231, and 243 DAT. Control of Leveillula taurica on living covers and Botrytis cinerea on rocoto was achieved with PROTEXIN^® 500 FW (carbendazim; 30 mL per 20 L; applied at 98, 112, and 180 DAT) and EPICO^® 750 WG (tebuconazole + azoxystrobin; 20 g per 20 L; applied at 148 DAT). For the control of Phytophthora capsici, PREDOSTAR^® (propamocarb + metalaxyl; 30 g per 20 L; applied at 125 and 180 DAT) and ALIETTE^® WG (fosetyl-Al; 2.5 kg ha⁻¹; applied at 169, 186, and 208 DAT) were used. All products were applied via foliar spray, except for fosetyl-Al, which was applied through the irrigation system.

Harvesting was staggered using hand scissors to prevent damage and maintain fruit quality. Fruits were harvested with their peduncles attached to extend post-harvest shelf life, depending on their stage of maturity. The final harvest took place at 316 DAT, marking the end of the experiment.

2.5 Water requirement

The irrigation program was initially designed based on historical data from the Majes Meteorological Station. Subsequently, specific adjustments were made for each treatment using a soil moisture meter (FIELDSCOUT TDR 150, Spectrum Technologies, Inc., Aurora, Illinois, USA). The adjustments aimed to maintain a soil moisture content of around 9%, as Ratliff et al. (1983) reported that the typical moisture thresholds for loamy sand soil are 16% at field capacity and 7% at the permanent wilting point, representing 9% available water.

The water requirement was determined using the irrigation depth (ID), based on the formula proposed by Allen et al. (1998) , as shown in Equations 1, 2:

\begin{array}{l} I D = E T_{c} - P_{e} & (1) \end{array}

\begin{array}{l} E T_{c} = E T_{0} \times K_{c} & (2) \end{array}

Where $I D$ is the irrigation depht (mm), $E T_{c}$ is the crop evapotranspiration (mm), $P_{e}$ is the effective precipitation (mm), $E T_{0}$ is the reference evapotranspiration (mm), and $K_{c}$ is the crop coefficient. $P_{e}$ was estimated using the method proposed by the USDA (Smith, 1992). The $E T_{0}$ value was calculated using the Penman-Monteith equation modified by the FAO (Allen et al., 1998), based on climate data recorded at the Santa Rita weather station. The $K_{c}$ coefficient varied according to the crop’s phenological stage, with the following values: 0.65 during vegetative growth, 0.97 at flowering, 1.06 during fruiting, and 1.13 at maturity (Sánchez Vesga et al., 2003).

2.6 Establishment of coverage

The selected species were Trifolium pratense var. americano, whose seeds were obtained from the local market, and Melilotus albus, whose seeds were collected from the field due to its behavior as a weed. Although the introduction pathway of M. albus into Peru is unknown, it has successfully adapted to desert regions such as Arequipa (Quipuscoa Silvestre et al., 2016) and is currently naturalized in various agricultural systems. Propagation was conducted in 288-cell trays using PROMIX-GTX (a blend of blond peat and vermiculite) as the substrate. Seven seeds were sown per cell. At 125 DAT of the rocoto crop, M. albus and T. pratense seedlings, 35 days old, were transplanted according to their respective T1 and T2 treatment layouts. The seedlings were arranged in four rows (two on each side of the main crop) with 10 cm spacing between plants, following manual weeding to ensure uniform experimental conditions. Treatment T3 involved a synthetic cover consisting of white-on-black plastic mulch (50-micron polyethylene with >60% visible light reflectance), which was installed at 28 DAT. Treatment T4 consisted of a no-cover control, where manual weeding was performed on 44, 61, 98, 125, and 155 DAT.

2.7 Biometric and physiological parameters

Plant height was measured every 7 days, starting at 5 DAT, in six representative plants per experimental unit, recording the distance from the stem base (collar) to the apex of the plant stem. The chlorophyll index was assessed in fully developed young leaves following the methodology described by Hu et al. (2010). Three plants per experimental unit were evaluated, considering three readings taken per plant. Measurements were conducted using a SPAD-502 Plus chlorophyll meter (Soil Plant Analysis Development), which provides an index proportional to the chlorophyll content in the leaf. Evaluations of the plants’ phenological behavior included monitoring the onset of flowering. An experimental unit was considered to have reached this stage when more than 50% of the plants exhibited visible signs of flowering (Arcila Pulgarín, 2007).

After harvesting, the fruits were selected and classified by size and weight according to the criteria described by Sardón Mamani (2015) Bola, with a height of 5.1 cm, width of 5.6 cm, and weight less than 100 g; Primera, with a height of 6.2 cm, width of 6.8 cm, and weight between 100 and 140 g; Extra, with a height of 6.9 cm, width of 8.5 cm, and weight between 140 and 160 g; and Super Extra, with a height of 8.7 cm, width of 9.6 cm, and weight greater than 189 g. Classification was performed on six representative plants per experimental unit, from which yield data were obtained. Yields were extrapolated to tons per hectare (t·ha¹) and number of fruits per treatment. The dry matter content was determined using a 5 g fruit sample, which was cut into cubes, homogenized, and weighed with an electronic balance (Explorer™ Pro Precision, Ohaus, USA; accuracy: 0.001 g). The sample was dried in an oven (Memmert Schwabach 854) at 105°C until a constant weight was achieved, following the AOAC 930.15 method (Brilhante et al., 2024).

Soil volumetric moisture was monitored using a FIELDSCOUT TDR 150 (Time Domain Reflectometer), with six measurements taken per experimental unit every 7 days starting at 33 DAT, at a depth of 20 cm. This depth was selected based on Grasso et al (Grasso et al., 2020), who reported that under drip irrigation, approximately 80% of the root system of Capsicum annuum is concentrated within the top 20 cm of soil. Additionally, it has been established that the easily evaporable layer in sandy loam soils is approximately 5 cm deep (Lehmann and Or, 2024) and that the wetting bulb depth in drip irrigation systems reaches around 20 cm in soils with similar characteristics under comparable irrigation durations (Hao et al., 2007).

2.8 Determination of weed control and living cover biomass

Weed control by living covers was assessed by counting the number of individuals per square meter (weed density), as described by Braun-Blanquet (1932). Three representative evaluations were conducted at 173, 202, and 254 DAT. Weed species were classified into functional groups based on their morphological characteristics: broadleaf weeds (dicotyledons) and narrow-leaf weeds (grasses), according to Najul and Anzalone (2006).

To evaluate the biomass of the living covers (Melilotus and Trifolium), three samples were collected per experimental unit following the method described by Baldivieso et al (Baldivieso-Freitas et al., 2018). Each sample covered an area of 1 m² and was taken from the head, middle, and tail sections of different beds. The aerial parts of the plants were cut and weighed to determine fresh weight, while the roots were extracted and washed before being weighed. Both the aerial and root parts were dried in an oven at 60 °C for 48 hours to obtain their dry weights. Dry biomass was then calculated based on these measurements.

2.9 Determination of N, P, and K in the biomass of living covers

To determine the total nitrogen (N), phosphorus (P), and potassium (K) content in the aerial and root biomass of the living covers, representative subsamples were collected, oven-dried at 65°C to a constant weight, ground, and sent to the laboratory (150 g per experimental unit). Total nitrogen was quantified using the Kjeldahl method, following the ISO 11261:1995 (International Organization for Standardization, 1995). Total phosphorus was determined using the AS-11 method, based on the Bray and Kurtz procedure (Bray and Kurtz, 1945), as specified in the Mexican Official Standard NOM-021-RECNAT-2000 (DOF - Diario Oficial de la Federación). Total potassium was analyzed using the modified AS-12.2002 method, as described by Semarnat et al (DOF - Diario Oficial de la Federación). All analyses were conducted at the Soil, Water and Foliar Analysis Laboratory (LABSAF) of INIA Arequipa.

2.10 Statistical analysis

Variables assessed at the end of the growing season (i.e., fresh and dry yields, and dry biomass and nutritional composition of living mulches) were evaluated using analysis of variance (ANOVA). Significant differences were analyzed using Tukey’s test (α = 0.05). Assumptions of normality and homoscedasticity were verified using Shapiro-Wilk and Bartlett’s tests, respectively.

Variables measured repeatedly over time (i.e., plant height, chlorophyll index, soil volumetric moisture content, and weed density) were analyzed using generalized linear mixed models (GLMM), also evaluating the random effect of plots. The corrected Akaike information criterion (AICc) was applied for model selection. Thus, Tweedie distribution with square root link was used for plant height and SPAD variables; beta distribution with logit link for soil moisture; and Tweedie distribution with logarithmic link for weed density. The significance of fixed effects and their interactions was evaluated using likelihood ratio tests (Chi²), and statistically significant factors and combinations (p < 0.05) were reported.

Discrete variables (i.e., flowering onset and number of fruits per plant) were analyzed using the Kruskal-Wallis test, applying Dunn’s test as post hoc analysis (α = 0.05) in case of significant differences. InfoStat version 2020 (Di Rienzo et al., 2020) was used for ANOVA analyses, and R version 4.4.3 (R: The R Project for Statistical Computing) for the other aforementioned analyses.

3 Results

3.1 Development of living covers

Trifolium pratense germinated 3 days after sowing (DAS), while Melilotus albus emerged at 7 DAS. During the seedling stage, T. pratense exhibited greater uniformity and vigor, with an average of 6 viable seedlings per cell, compared to only two seedlings in M. albus, which was affected by intraspecific competition. At the time of transplanting (at 35 DAS), T. pratense reached a height of 12 cm and displayed dense foliage, whereas M. albus measured 8 cm and exhibited limited foliar development. In field conditions, however, M. albus demonstrated accelerated growth, attaining a height of 135 cm at 132 days, in contrast to 30 cm for T. pratense. In terms of root development, T. pratense formed a dense fasciculated and shallow root system, while M. albus developed a thick taproot with lower lateral root density.

3.2 Biometric and physiological variables

The GLMM used for plant height explained 98% of the observed variation in the data (Supplementary Table S1). Thus, plant height (Supplementary Figure S1) varied significantly as a function of time (i.e., days after transplanting), but not according to the applied treatments. Regarding the chlorophyll index, the best-fitting model explained no more than 37% of the observed deviance in the data (Supplementary Table S2). Overlapping confidence bands were identified for the five evaluation timepoints across the four treatments (Supplementary Table S3). They highlight the non-significant differences in the evaluations conducted after the implementation of the cover crops.

3.3. Yield variables

No significant differences were observed in fruit number per plant, fruit dry matter (Supplementary Table S4), or total yield (Table 1). However, T3 and T4 exhibited a tendency toward higher values with reduced variability. The analysis by commercial categories revealed significant differences for Primera Category (C3), where T4 achieved the highest values. No treatment produced fruits classified as Extra Super Category (C1).

Table 1

Table 1. Rocoto yield by commercial categories.

3.4 Soil volumetric moisture

The GLMM with beta distribution and logit link explained 95% of the variation in soil volumetric moisture (SVM) (Supplementary Table S5). A significant effect was observed in the quadratic interaction between treatments and time, revealing differentiated trajectories. T1 and T4 maintained stable moisture levels (≈12–15%) throughout the entire period, while T2 and T3 exhibited contrasting patterns. T2 started with the lowest SVM (≈11.1%) and increased progressively toward day 245, whereas T3 initiated with the highest moisture content (≈28.9%) and decreased quadratically after the plastic mulch was installed and the irrigation time proportion was reduced by approximately 50% (Figure 3B). The establishment of living coverage in T1 and T2 increased their irrigation time proportion (Figure 3A). After 166 DAT, the volume of water applied to T1 and T2 was increased. T2 significantly showed greater soil moisture content than T1 despite both receiving the same irrigation volumes. The maximum effect was recorded at 257 DAT (maturation of M. albus) coinciding with the peak development of the aerial and root biomass of the living cover. From 266 DAT onward, all treatments converge toward similar irrigation time proportions, coinciding with the crop’s maturation stage.

Figure 3

Graph A shows irrigation time distribution as a stacked area chart with shading in beige, gray, pink, and green. Graph B displays soil moisture percentage over days after transplanting with four modeled curves in black, green, pink, and brown. Red dashed lines mark specific days, such as T3 – Pi (28 DAT) and T1 – LCm (257 DAT). Equations and symbols are included for explanation.

Figure 3. Irrigation time distribution (%) and soil moisture (%). (A) Proportional distribution of irrigation time among treatments throughout the crop cycle. (B) Evolution of soil moisture by treatment during the cropping period. Arrows indicate key events: 28 DAT – Plastic installation (Pi) in T3; 33 DAT – Irrigation reduction (Ir) in T3; 125 DAT – Living cover transplanting (LCt) in T1 and T2; 163 DAT – Irrigation increment (Ii) in T1 and T2; 166 DAT – Living cover at 90% (LC90%) in T1 and T2; 257 DAT – Living cover maturity (LCm) in T1.

Analysis of variance revealed significant differences among treatments for irrigation water productivity (IWPt) (F = 8.53; p = 0.0139). Treatment T3 demonstrated the highest water use efficiency, with IWPt values significantly superior (2.2 and 2.0 times) compared to T1 and T2, respectively (Table 2). Treatment T4 exhibited intermediate behavior, with no significant differences from T3 or the T1-T2 group. These results reflect a significant effect of applied irrigation volume on IWPt, where the volume reduction in T3 (6690 m³ ha⁻¹) resulted in statistically higher efficiency compared to treatments with greater water input (T1 and T2: 9635 m³ ha⁻¹).

Table 2

Table 2. Irrigation Water Productivity (kg m^–3) related to the total rocoto yield.

3.5 Potential contribution of biomass and nutrients to the soil

T1 produced significantly greater above-ground dry biomass (8.57 t·ha⁻¹) compared to T2 (3.04t·ha⁻¹), whereas T2 exhibited higher root biomass (1.42 vs. 0.56 t·h⁻¹) as shown in Figure 4. The total N, P, and K content potentially contributed to the soil by living covers did not differ significantly between treatments (Table 3). However, organ-specific analysis revealed distinct patterns: Trifolium pratense (T2) showed higher concentrations of P and K in the roots, while Melilotus albus (T1) was notable for its elevated K content in the above-ground biomass.

Figure 4

Bar chart comparing dry biomass under two treatments, T1 and T2, displayed in green and pink. Categories include Aboveground Biomass, Root Biomass, and Total. T1 shows higher aboveground biomass than T2, but less root biomass than T2. Consequently, T1 total dry biomass is higher than T2, with significant differences marked by letters a and b.

Figure 4. Organ-specific dry biomass of Melilotus albus (T1) and Trifolium pratense (T2), separated into aboveground biomass, root biomass, and total biomass. Data are presented as mean ± standard error. Different letters above the bars indicate statistically significant differences between treatments for each evaluated component (Tukey test, p < 0.05).

Table 3

Table 3. Contribution of N, P, and K from living cover biomass.

3.6 Control of broadleaf weeds and grasses

The GLMM model (Tweedie distribution, log link; AICc = 216.16) explained 99% of the variation in weed density. The Treatment × Type interaction (χ² = 9.69; p < 0.01) was significant (Supplementary Table S6), as was the main effect of time. Weed presence decreased progressively over time (Figure 5A); T1 and T2 showed similar efficacy in controlling broadleaf weeds, whereas T2 demonstrated superior efficiency in controlling grass-type weeds (Figure 5B).

Figure 5

Bar chart showing plant density in two segments. Chart A displays density over three days after transplanting: 173, 202, and 254, with the highest density at day 173. Chart B shows density for the combinations of treatments and weed types: T1 Broadleaf, T1 Grasses, T2 Broadleaf, and T2 Grasses, with T2 Broadleaf having the highest density. Error bars indicate variability, marked by letters a, b, and c for significance.

Figure 5. Density of broadleaf species and grasses. (A) Main effect of DAT. (B) Effect of treatment × weed interaction. Error bars represent 95% confidence intervals. Different letters indicate significant differences in density at a significance level of 0.05 according to the False Discovery Rate (FDR) correction method.

4 Discussion

4.1 Uniform behavior among treatments

The cover treatments (T1: Melilotus albus, T2: Trifolium pratense, T3: plastic mulch, T4: no cover) did not result in significant differences in plant height, Chlorophyll index, flowering onset, fruit number, or dry matter content of rocoto fruits. This lack of variation could be attributed to factors related to nitrogen nutrition and the timing of evaluation. Aminifard et al. (2012) demonstrated that variations in nitrogen fertilization significantly influence these parameters in peppers; however, in the present study, all treatments received the same mineral fertilization regime. As noted by Havlin et al. (2016), symbiotically fixed nitrogen is primarily utilized by the host legume, with minimal release into the soil during its life cycle. It becomes available to the main crop only after the legume biomass is decomposed and incorporated into the soil.

Chu et al. (2017) emphasize that the evaluation period is a determining factor, as the benefits of living covers tend to manifest predominantly in the medium- to long-term. Similarly, plastic mulch did not exhibit immediate effects. Hochmuth and Hochmuth (1994) reported that white-on-black plastic mulch did not influence the number of pepper fruits during the first year of evaluation. However, positive effects emerged in the second year.

4.2 Response of fruit yield to cover management

In this study, no statistically significant differences were detected in total yield (Y) or in the Extra (C2) and Bola (C4) fruit categories among treatments. Although certain trends were observed in mean values, these should be interpreted with caution due to the high variability recorded, particularly in treatment T2. This level of dispersion reduced the statistical power of the analysis and limited the detection of effects attributable to cover management, a situation also reported by other authors in experiments with high genetic or environmental heterogeneity (Gómez and Gomez, 1984).

These results are consistent with studies in Capsicum and other crops showing that traits related to yield and quality are highly sensitive to environmental factors such as radiation, humidity, and nutrient availability (Zewdie and Bosland, 2000; Sahmat et al., 2024). Consequently, genotype stability cannot be evaluated solely on the basis of mean yield, but requires considering its differential response under contrasting environmental conditions (Sran et al., 2021; Sahmat et al., 2024). A key aspect to consider is the marked climatic discrepancy between the origin of the genetic material and the experimental site. The Oxapampa accession is associated with a production zone characterized by high relative humidity (88%) and low solar radiation (166 W·m⁻²). In contrast, Santa Rita de Siguas presents opposite conditions, with 49% relative humidity and 279 W·m⁻² solar radiation, according to SENAMHI data (SENAMHI - Descarga de Datos). This genotype × environment mismatch likely constrained the expression of genetic potential, regardless of treatment. In this regard, climate represents a particularly complex factor, as it integrates relatively predictable components defined by the prevailing climatic regime with unpredictable components derived from interannual variability (Oladosu et al., 2017). Moreover, the relative stability of fruit load, in contrast with the variability of quality traits, highlights the complexity of genotype × environment interactions and their influence on yield stability.

In the Primera (C3) category, significant differences were detected, with treatment T4 (manual weeding) achieving the highest yield. Its superiority over living covers (T1 and T2) may be related to the reduction of interspecific competition, while the difference with plastic mulch (T3) could be associated with accelerated fruit ripening, as previously reported in studies with white mulch (Overbeck et al., 2013). This earlier ripening may have favored the harvest of smaller fruits, thereby reducing the proportion of C3 fruits. However, under conditions of high radiation and low humidity, the influence of both environment and genetic material on this differential response cannot be ruled out. Since this study was conducted during a single growing season, the conclusions should be interpreted within these limitations, as they do not account for interannual variability that could substantially modify the genotype’s response.

4.3 Dynamics of volumetric soil moisture

Optimizing water resources is essential in arid regions. In this context, plastic mulch (T3) proved highly effective in conserving water, reducing irrigation time by about 50% and maintaining soil moisture at an optimal content for plant-available water in sandy loam soils (Ratliff et al., 1983). These results align with those of Díaz-Pérez (2010) and Ren et al (Ren et al., 2021), who found higher water content in the top 60 cm of soil when using plastic mulch. Among living cover treatments, T2 shows greater soil moisture retention than T1, indicating improved water use efficiency once fully established. This advantage of T. pratense in conserving water is likely linked to differences in its root system. M. albus develops deeper, less dense roots near the surface (Ríos et al., 1993), while T. pratense has a denser, more superficial root system (Fergus and Hollowell, 1960; Bowley et al., 1984), which favors moisture retention in the upper soil layers. High root density decreases macroporosity (Daly et al., 2015), possibly limiting rapid infiltration but increasing water retention capacity. Additionally, root exudates can alter soil structure and absorb large amounts of water (Xiao et al., 2024), boosting water availability in the rhizosphere. Morphologically, M. albus grows upright and taller, resulting in a larger leaf area exposed to sunlight, which leads to higher transpiration rates. Conversely, the prostrate growth habit and shorter stature of T. pratense keep its foliage shaded beneath the rocoto, reducing water loss through transpiration.

The literature confirms that bare soil promotes greater water loss through direct evaporation (Brun et al., 1986; Gomes de Andrade et al., 2011), whereas living covers help reduce both evaporation and runoff, thereby enhancing infiltration and water conservation (Wyngaarden et al., 2015). However, as we found, it depends on the living covers’ characteristics. Biazzo and Masiunas (2000) reported that living mulch with T. pratense retains more moisture compared to bare soil. Nevertheless, Nielsen and Vigil (2005) and Blanco-Canqui et al. (2015) cautioned that in semi-arid environments, living covers can reduce the moisture available to the main crop by 20–50%, depending on the timing of their suppression and the management practices applied.

To the best of our knowledge, no previous reports of IWPt values are available for Capsicum pubescens. In our study, IWPt ranged from 4.11 to 9.17 kg m⁻³, with higher values observed under plastic cover. However, the variability in the yields obtained, compared with those reported in the literature (Pérez-Grajales et al., 2004; Puente et al., 2014; MIDAGRI, 2025), suggests that further research is needed to better understand water productivity responses in this crop.

4.4 Biomass production by living plant covers and nutrient accumulation

In terms of biomass contribution, T1 (Melilotus albus) produced 2.05 times more dry above-ground biomass than T2 (Trifolium pratense). This difference is attributed to the erect growth habit and greater height of the Melilotus genus, which enables more effective light interception and higher photosynthetic efficiency (Fontana et al., 2018), further supported by its rapid vegetative regrowth from basal shoots (Țîței, 2022).

Conversely, T2 (Trifolium pratense) exhibited 2.6 times greater dry root biomass, consistent with its capacity to develop dense and deep root systems (Blanco-Canqui et al., 2015; Dlamini et al., 2024), with reported root dry matter yields ranging from 2.3 to 2.9 t·ha^-¹.

The above-ground to root biomass ratio differed significantly between species These proportions reflect contrasting ecophysiological strategies. Melilotus albus prioritizes above-ground biomass, which is advantageous for forage production or rapid surface coverage, (Țîței, 2022). In contrast, Trifolium pratense allocates more biomass to its root system, making it more suitable for biopore formation, enhanced water infiltration and retention, and increased rhizospheric biological activity (Yu et al., 2016; Mckenna et al., 2018; Adetunji et al., 2020 ).

Living covers can generate significant agronomic interactions with the main crop. In this context, the high above-ground biomass production of Melilotus albus may increase competition for light in high-density systems, potentially affecting the development of rocoto pepper. White et al. (2016) note that fast-growing covers with early foliage closure can exert intense physical competition and may also release allelopathic compounds. In contrast, species such as Trifolium pratense, which produce less above-ground biomass, may reduce such competition in polyculture systems (Thorsted et al., 2006). Their greater biomass allocation to the root system, however, contributes to improved soil physical structure and the formation of stable aggregates (Mckenna et al., 2018).

As leguminous species, both Melilotus albus and Trifolium pratense contribute to biological nitrogen fixation; however, the total amount fixed may vary due to differences in biomass production (McEwen and Johnston, 1985; Nesheim and Øyen, 1994; Hernandez Escareño, 1998; Pederson et al., 2002; Terroba et al., 2024).In the present study, although the differences were not statistically significant. This variability among studies can be explained by differences in nitrogen fixation efficiency among legume species, which is influenced by nodule morphology and root system development, with higher efficiency observed in species with more developed root systems (Terpolilli et al., 2012).

The total nutrient accumulation (from root and above-ground biomass) indicates that Melilotus albus (T1) can contribute 78.09 kg N·ha^-¹, 25.05 kg P·ha^-¹, and 130.64 kg K·ha^-¹. In comparison, Trifolium pratense (T2) contributes 47.00 kg N·ha⁻¹, 16.08 kg P·ha⁻¹, and 91.97 kg K·ha⁻¹. However, this contribution will only effectively improve soil fertility after biomass incorporation, and the availability of nutrients for the subsequent crop will depend on the decomposition rate, moisture and temperature conditions, the residue’s C/N ratio, and soil microbial activity (Havlin et al., 2016; Weil and Brady, 2017).

4.5 Broadleaf and grass weed control

Although treatments T3 (plastic mulch) and T4 (manual weeding) are not shown in Figure 5, both achieved complete weed control. The synthetic cover and manual weeding eliminated weed competition, aligning with findings by Kasirajan and Ngouajio (Sokombela et al., 2025), who reported that plastic mulches block photosynthetically active radiation, which is necessary for weed germination. Similarly, studies by Barla et al. (2018) and Mzabri et al. (2021) confirm that plastic mulching effectively reduces weed biomass, yielding results comparable to those achieved with chemical treatments and superior to those obtained with organic mulching.

Treatments with living covers (T1 and T2) showed no statistically significant differences in terms of broadleaf weeds; however, T2 was more effective at controlling grasses weeds, suggesting a selective suppressive effect. Living covers suppress weeds through multiple mechanisms: allelopathy (Fujii, 2003), competition for resources (Williams et al., 1998), reduced light availability (Teasdale and Mohler, 2000), and the release of allelochemical compounds that enhance shading and competition (Mohammadi and Mohammadi, 2012).

The temporal weed control T2 suggests that this cover may favour the growth of particular grass species. According to Brainard et al (Brainard et al., 2011), leguminous covers may be less effective at weed suppression due to unpredictable fluctuations in temperature and soil fertility, reducing their suppressing effectiveness.

The results have practical implications for integrated weed management. While synthetic mulches and manual weeding offer higher effectiveness, they require greater financial investment or labour inputs (Steinmetz et al., 2016). In contrast, plant covers, although less effective at suppressing weeds, provide additional benefits, including moisture conservation, contribution of organic matter, and enhanced biodiversity (Teasdale, 1996; Blanco-Canqui et al., 2015). However, the prolonged use of plastic mulch can alter soil biocoenosis, increase greenhouse gas emissions, and lead to the accumulation of plastic residues in the soil (Steinmetz et al., 2016). Although biodegradable alternatives are available, their high cost limits their widespread adoption compared to polyethylene mulch (Sokombela et al., 2025), underscoring the need for a comprehensive assessment of their use. This challenge is exemplified in China, where the accumulation of 550,800 tons of plastic residues resulted in a 6–10% reduction in cotton yields (Zhang et al., 2020). The problem is further exacerbated by the projected growth of the global synthetic mulching market, estimated to reach USD 19.49 billion by 2029 with an annual growth rate of 7.1% (MarketsandMarkets, 2025), while the high recovery costs undermine its long-term economic feasibility (Steinmetz et al., 2016). On the other hand, manual weeding, although effective for weed removal, has limitations that reduce its feasibility in large areas due to its high demand for labor, physical effort, and time, which substantially increase operational costs (Hujerová et al., 2016; Assani Bin Lukangila et al., 2024). Moreover, its environmental benefits are limited compared to those provided by living covers (Poeplau and Don, 2015; Koudahe et al., 2022).

4.6 Management recommendations for cover crops in arid zones

4.6.1 Scenarios of severe water scarcity

In the context of the arid regions of Arequipa, where water availability is limited, plastic mulch showed potential as a strategy to conserve soil moisture and reduce weed incidence, which could contribute to improving water use efficiency (WUE).

As a sustainable alternative, it is recommended that the economic viability of biodegradable plastics be assessed. Despite their higher initial cost, their implementation should be evaluated through a cost-benefit analysis that considers reduced labor requirements for post-harvest removal, decreased environmental impact, and the potential for enhanced marketability of products under sustainable certifications.

4.6.2 Strategies for organic matter enrichment

Arid soils, such as those found in Arequipa, with low organic matter content (1.31% in this study), tend to present limitations in water retention, cation exchange capacity, and biological activity. In this context, high-biomass leguminous cover crops, particularly Melilotus albus due to its drought tolerance, nitrogen-fixing capacity, and high above-ground biomass production, may represent a promising alternative. The biomass could potentially be used as green manure or compost input, contributing to a circular economy approach within systems that also employ plastic mulch.

4.6.3 Management under moderate water availability conditions

In arid zones where water availability is sufficient, the establishment of Trifolium pratense may offer potential benefits, as it can provide multiple ecosystem services such as improved soil moisture retention, increased organic matter content, and atmospheric nitrogen fixation.

Furthermore, in compacted soils, its dense and fasciculated root system may contribute to biological decompaction, enhancing water infiltration and improving soil structure. This hypothesis requires experimental validation through medium- and long-term monitoring of soil physical parameters.

4.7 Study limitations and recommendations for future research

The main limitations of this study include its confinement to a single agricultural season, which restricts the assessment of interannual variability and the cumulative effects of living covers over the medium and long term. The limited number of replicates reduces the statistical robustness of the findings and hampers the detection of significant differences among treatments. Furthermore, variations in irrigation timing, implemented as part of the management strategy, prevented valid statistical comparisons of soil moisture retention across all treatments, thereby limiting the objective quantification of this key parameter. Finally, measuring moisture only at 20 cm depth may not fully capture the dynamics in deeper soil layers.

For future research, it is recommended to integrate management variables, soil moisture measurement at different depths, differentiated irrigation levels, crop associations, and various cover crop species (including grasses, cruciferous plants, and legumes) according to the local agroecological conditions, and family farming systems.

5 Conclusions

The use of soil covers in rocoto pepper (Capsicum pubescens) cultivation under arid conditions was insufficient to produce significant effects on most production variables evaluated during the growing season studied. However, plastic mulching showed potential for water use efficiency and weed control under the established experimental conditions. Among the living mulches evaluated, Trifolium pratense tended to register greater soil moisture retention than Melilotus albus, with differences of up to 63.3% observed during the maturation stage of Melilotus albus. In terms of biomass contribution to the soil, Melilotus albus evidenced favorable characteristics as a living mulch contributor. This behavior could be associated with its morphological characteristic and observed growth patterns. Regarding nitrogen input through biomass incorporation, both leguminous species demonstrated nitrogen accumulation potential, with Melilotus albus exhibiting numerically higher values that did not achieve statistical significance (P > 0.05). For weed control, both living mulches provided moderate weed suppression without significant differences between species, though showing reduced efficacy compared to plastic mulching or manual weeding. The patterns observed in this study require multi-year evaluation to establish their consistency and validity. Consequently, future research should incorporate medium to long-term experimental designs, increase replication numbers, and broaden the scope of assessed variables to derive more robust conclusions regarding soil cover utilization in rocoto pepper production under arid conditions.

Data availability statement

The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.

Author contributions

RP: Formal Analysis, Investigation, Methodology, Writing – original draft. EC: Conceptualization, Formal Analysis, Investigation, Methodology, Writing – original draft. RF: Conceptualization, Writing – review & editing. AQ: Methodology, Writing – original draft. RS: Supervision, Writing – review & editing.

Funding

The author(s) declare that financial support was received for the research and/or publication of this article. This research was funded by the INIA project CUI 2487112 “Mejoramiento de los servicios de investigación y transferencia tecnológica en el manejo y recuperación de suelos agrícolas degradados y aguas para riego en la pequeña y mediana agricultura en los departamentos de Lima, Áncash, San Martín, Cajamarca, Lambayeque, Junín, Ayacucho, Arequipa, Puno y Ucayali”.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declare that no Generative AI was used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fagro.2025.1663633/full#supplementary-material

References

Adetunji A. T., Ncube B., Mulidzi R., and Lewu F. B. (2020). Management impact and benefit of cover crops on soil quality: A review. Soil Tillage Res. 204, 104717. doi: 10.1016/j.still.2020.104717