Microclimatic Parameters, Soil Quality, and Crop Performance of Lettuce, Pepper, and Chili Pepper as Affected by Modified Growing Conditions in a Photovoltaic Plant: A Case Study in the Puglia Region (Italy)

Abstract

The performance of lettuce, pepper, and chili pepper, and the biological soil quality, in a ground-mounted PV system under cultivation conditions typical of the Mediterranean environment of the Puglia region were evaluated. Microclimatic parameters, plant growth and yield response, soil quality assessed using the QBS-ar index, and land equivalent ratio (LER) were determined in three different cultivation areas: a cultivation area outside the photovoltaic plant but immediately adjacent to it (‘Control’); the inter-row area closest to the row of panels exposed to sunlight (‘Area close PV structure’); the inter-row area distant from the row of panels (‘Area distant PV structure’). Cumulated solar radiation, in particular during the summer growing cycles, was only slightly affected in the Area distant PV structure (1616 and 2130 MJ m⁻² for pepper and chili pepper, respectively, in the control area, in comparison to 1630 and 2044 MJ m⁻², in the Area distant PV structure), while it was strongly reduced in the Area close PV structure (883 and 1091 MJ m⁻² for pepper and chili pepper, respectively). In general, a reduction in air temperature and wind speed, as well as an increase in relative air humidity, was observed under PV conditions. On average, the evapotranspirative demand was reduced in the PV growing conditions compared to open field, with a more relevant effect in the sub-zone close to the photovoltaic structures, where cumulative ET0 was 28% and 34% lower than the Control in the pepper and chili pepper growing cycle, respectively. Lettuce growth was impaired by PV cultivation conditions, with an average reduction of 15% in plant height and 37% in marketable yield per plant, with no significant differences between the two sub-zones in the PV system. For pepper, the best growing conditions were observed in open field control compared to PV, but with differences related to the PV sub-zone. The plants grown in the Area distant PV structure were more negatively affected by the modified growing conditions, showing the lowest shoot and fruit fresh weight, the latter reduced by 51% compared to the Control; intermediate values were observed for these parameters in the Area close PV structure, with a less severe tendency to yield reduction. For chili pepper, both shoot and fruit fresh weight were lower in PV conditions, regardless of the sub-zone, with a reduction of 82% in yield per plant compared to the Control. However, despite the yield reductions, the LER was improved (1.60 and 1.40 in case of a lettuce + pepper or lettuce + chili pepper annual cropping program, respectively), highlighting a more efficient use of land, without negative or even ameliorative impacts on biological soil quality and biodiversity in terms of QBS-ar and microarthropods taxa abundance. Knowledge of the response of different crops under cultivation conditions typical of specific environments is necessary to define optimal cropping programs aimed at maximizing resource-use efficiency and land use.

1. Introduction

Increased demand for food and energy at the global level implies problematic competition for agricultural land. Recently, the interest toward the integration of agricultural and energy production by photovoltaic (PV) systems known as agrivoltaics (AV) has been growing worldwide, in principle representing a promising approach to face the challenges of food security and energy supply from renewable sources [1].

In Italy, as well as in several other countries, the possibility of large-scale implementation of AV systems is sparking heated debate and discussions at various levels with focus on potential benefits and drawbacks that merit careful consideration. In general, on one hand, combining agricultural production with photovoltaic energy generation can enhance land use efficiency and contribute to reach renewable energy and decarbonization targets; however, high costs, regulatory challenges, and the impacts on agricultural performances, soil quality conservation, and land use need to be addressed.

Recent studies focused on Italy as a case study reported that by allowing simultaneous agricultural and energy production, AV can significantly enhance land productivity with minimal land-use impact [2]. According to a spatial multicriteria land eligibility analysis, based on factors that can affect both the photovoltaic potential and the agricultural use, it is estimated that an AV coverage of 1.24% of the eligible area would enable reaching the national 80 GW target of new renewable capacity to achieve the country’s decarbonization and energy transition objectives by 2030 [3]. The potential of AV in stimulating local economies by reducing energy costs for farmers and creating new job opportunities in the renewable energy sector is also outlined [4]. However, drawbacks and critical issues are reported as well. The initial investment for AV systems is significantly higher than traditional photovoltaic setups, which may deter adoption without supportive policies [5]. The procedures for approving projects and granting permits to build new installations are still subject to regulations, which are often too complex and ambiguous, aimed at protecting agricultural activity and landscapes [4]. Further doubts have been raised on the potential negative effects resulting from the transformation of traditional agro-ecosystems into AV systems, with specific reference to agro-biodiversity loss [2].

Although the most recent trend consists of focusing the concept of AV to systems characterized by innovative layouts, i.e., PV structures lifted off the ground and further adapted to meet the requirements of sufficient crop production underneath [6], several scientific reports dealing with crop cultivation between ground-mounted PV rows designate such systems as agrivoltaic [7,8]. The definition of the AV concept may also be subjected to the specific regulatory framework of the different territories. Based on official guidelines at the Italian national level [9], under the definition of AV, the systems are described as follows: (i) designed and built in such a way as to adopt a spatial configuration and appropriate technological choices, such as to allow the integration of agricultural activity and electricity production and enhance the production potential of both sub-systems; (ii) operated, during their technical life, in such a way as to guarantee the synergistic production of electricity and agricultural products and not compromise the continuity of agricultural and/or pastoral activity; (iii) provided with systems for monitoring the continuity of agricultural activity, namely the impact on crops and the continuity of the activities of the agricultural companies involved. The systems defined as “advanced AV systems” are those that meet additional requirements, such as the adoption of innovative integrated solutions with modules raised from the ground, and the implementation of systems aimed at monitoring the effects on water saving, soil fertility recovery, and climate change resilience.

Despite the increasing interest toward innovative structures and the perspective of new advanced installations, thousands of hectares of standard photovoltaic ground mounted and/or fixed configuration plants are still worldwide operative, with about 16,000 ha in Italy, suggesting the opportunity of exploring the implications of combining energy and agricultural production in such conditions [10].

From an agronomic point of view, in order to ensure the maintenance of profitable agricultural activity in PV plants, the main concern is related to the effects of modified microclimatic conditions (i.e., shading, air movement, temperature, and air relative humidity) on crop performance. Results on the effects of PV on crop yield are often contrasting. Although some studies indicate that agrivoltaic systems can ameliorate microclimate at the crop level resulting in improved yield [11,12], numerous studies report yield reduction, ranging from 3% to 62%, depending on the crop and panel configuration [13].

It has been proposed that the development of AV systems, combining agricultural production with solar energy generation, offers promising perspectives for sustainable land management. While agronomic implications and energy outputs of these systems have been the object of extensive research, their impacts on soil ecosystems remain poorly documented. Soil invertebrates, particularly microarthropods, play a key role in ecosystem processes such as organic matter decomposition, nutrient cycling, and soil structure maintenance. The QBS-ar index is an established tool for assessing soil biological quality, relying on the eco-morphological adaptations of microarthropods to edaphic environments. Each group is assigned a score (Eco-Morphological Index, EMI) ranging from 1 to 20, based on its degree of adaptation to soil life. The sum of the EMI scores of the taxonomic units present in a sample constitutes the final value of the QBS-ar index. QBS-ar provides an indication on soil biological quality related to land degradation, and is based on the morphological features mentioned above, assigning an Eco-Morphological Index (EMI), ranging between 1 and 20, in relation to the adaptation level to soil (1 = no adaptation; 20 = total adaptation) at each taxon. QBS-ar shows the results of the sum of each maximum EMI score assigned at each taxon identified in the soil sample (refer to Parisi et al., 2005 [14], and Menta et al., 2018 [15], for details of the method).

One of Italy’s most important agricultural regions is Puglia in the south-eastern part of the country, with about 80% of its total surface area used for agricultural purposes [16]. Although there is considerable unused and unproductive land in rural areas [17], the region has a leading position for vegetable crops production, as it accounts for 22% (70,340 ha) and 18% (1,413,304 tons) of open-field cultivated area and production at the national level [18]. A recent study estimated that, based on an analysis aimed to assess land use and potential AV opportunities, a minimal (namely 0.25% corresponding to 3250 hectares) occupation of utilized agricultural land would allow a regional installed AV capacity of 1.3 GW, considered a reasonable range of AV development in the region, which can contribute both to the energy transition and the support of the agricultural sector, especially in marginal areas [19].

Based on the above-mentioned considerations, the aim of this study was to evaluate the agronomic performance of different vegetable crops (lettuce, pepper, and chili pepper) and the biological soil quality, assessed using the QBS-ar index, in a ground-mounted PV system under cultivation conditions typical of the Mediterranean environment of the Puglia region. During the study, the microclimatic modifications of the growing environment were monitored with prototype and low-cost sensing and data transmission technologies specifically developed for agrivoltaic systems.

2. Materials and Methods

2.1. Site Description, Experimental Layout, and Growing Conditions

The study was conducted in a photovoltaic plant located in the Puglia region, Italy, in the Montemesola countryside, Taranto province, 40°32′32″ N, 17°20′18″ E, at an altitude of 69 m a.s.l. The system is composed of ground-mounted fixed photovoltaic modules with an orientation of approximately 78.8° with respect to North, i.e., slightly inclined towards the South-East with an almost East–West orientation. This configuration is optimized to ensure a more uniform distribution of solar radiation throughout the day, rather than maximizing production at midday, as occurs in South-oriented systems. The modules, 2.7 m high and inclined at 25°, are spaced 8 m apart.

The study area has a typical Mediterranean climate, with mild winters, and hot and dry summers. Minimum temperatures are usually recorded between January and February, but rarely fall below 5–7 °C. The warmest months are July and August, with average monthly temperatures ranging from 26 to 28 °C. The average annual rainfall is about 600 mm, concentrated in the autumn and winter. Relative humidity tends to be higher in the cold months (but rarely exceeds 80%), while in spring and summer it decreases, thanks to moderate winds from the northern quadrants, which favor ventilation and mitigate temperatures.

The soil has a loamy texture, good water retention and air capacity, normal salinity and medium alkaline reaction (EC 0.47 dS m⁻¹ and pH 8.1 in 1:2.5 water extract), high active limestone content (13.2%), and low organic matter (1.9 g 100 g⁻¹). The macro- and meso-elements are in a satisfying range: total nitrogen 1.18 g kg⁻¹, available phosphorus 21 mg kg⁻¹, exchangeable potassium, calcium, and magnesium 480, 2760, and 294 mg kg⁻¹, respectively. The cation exchange capacity is 17.67 meq 100 g⁻¹ and the base saturation percentage is 100%, indicating good potential soil fertility.

Three vegetable species were tested for their response to cultivation under PV conditions. The three growing cycles took place in 2024 in different seasons: lettuce (Lactuca sativa L. var. longifolia, cultivar F1 L7, Franchi Sementi, Milano, Italy; 16 February–7 May), pepper (Capsicum annum L., cultivar Corner F1, ISI Sementi, Fidenza, Italy; 5 June–27 August), and chili pepper (Capsicum annum L., cultivar Gorria F1, ESASEM, Casaleone, Italy; 5 June–10 October).

For the purposes of the study, three different cultivation areas were used and compared. In the photovoltaic field, the area between two rows of photovoltaic structures, excluding a 0.8 m strip immediately adjacent to the pillars of the structures on both sides, was virtually divided into two areas, the one closest to the row of panels exposed to sunlight, and therefore most affected by shading phenomena (‘Area close PV structure’), and the one further away, therefore less affected by shading (‘Area distant PV structure’). A cultivation area outside the photovoltaic plant but immediately adjacent to it was also included in the study (‘Control’).

For the three experiments, seedlings were obtained from a local nursery and transplanted (the lettuce on 16 February, and the pepper and chili pepper on 5 June) with a plant density of 6.2 plant·m⁻² (distance of 0.4 × 0.4 m) for lettuce, and 3.1 plant·m⁻² (distance of 0.4 between plants × 0.8 m between rows) for the pepper and chili pepper. In the summer experiments, a black 105 g·m⁻² UV-stabilized, water- and air-permeable woven polypropylene mulch sheet was used to control weeds and reduce evaporation.

Seedlings were transplanted in the three different cultivation areas. Irrigation was provided with drip tubes (2.2 L h⁻¹ drippers 0.5 m apart). Total irrigation volumes of 20, 90, and 109 mm were provided for lettuce, pepper, and chili pepper, respectively.

The AV plant was provided with an innovative system, SolarFerigation, designed and implemented using low-cost sensing and IoT technologies combined with advanced mechanical solutions for smart irrigation management and environmental parameter monitoring. The technical features of the innovative system implemented in a prototype version have been recently described in detail, and the results of tests for technological performances have been reported [20]. Briefly, the system uses autonomous photovoltaic technologies to power its operations, reducing energy costs and making the system particularly suitable to be used in AV plants, and usable in contexts where it is not possible to connect to the traditional power grid. The system is composed of two main components: the central unit, enabling precise distribution of water and fertilizers in irrigation zones, and the sensor nodes, which allow the setup of sensor networks for collecting environmental and soil data.

Based on average nutrient uptake and common fertilization practice for the crops in the area, 110 kg·ha⁻¹ N, 70 kg·ha⁻¹ P-P₂O₅, and 160 kg·ha⁻¹ K-K₂O for lettuce, and 200 kg·ha⁻¹ N, 90 kg·ha⁻¹ P-P₂O₅, and 275 kg·ha⁻¹ K-K₂O for pepper and chili pepper were applied via fertigation, dividing the total dose during the crop cycle according to the crop needs in the different phenological phases.

2.2. Measurements

2.2.1. Environmental Data/Microclimatic Variables

For the purposes of microclimate monitoring in the different cultivation areas under comparison, the SolarFertigation system was used to collect meteorological data. As reported above, the system allows the deployment of an IoT sensor network, the acquisition and the processing of the related data through LoRaWan technology and cloud-computing logic [20]. Please refer to the aforementioned publication for a detailed description of technological aspects. For the lettuce experiment, two IoT weather station nodes were installed at 1.50 m above ground: one in the control area and one in the center of an inter-row lane between the PV sturctures, respectively. For the pepper and chili pepper cycles, weather stations were increased, and nodes were installed in the control area, in the area under PV structures, and in the area distant from PV structures. The sensor nodes were equipped with sensors for environmental parameters: solar irradiation (VEML7700-TT, Vishay Intertechnology, Shelton, CT, USA), air temperature and relative humidity (DHT20, AZ-Delivery Vertriebs GmbH, Deggendorf, Germany), and, only for the pepper and chili pepper growing cycles, wind speed (CALT 9–30 V DC Supply 0–5 V Output 0–45 m s⁻¹ Range Air 3 Cup F1J, Shanghai QIYI Electrical & Mechanical Equipment Co., Ltd., Shanghai, China). Data were collected at a 15 min interval and then averaged or integrated at a daily scale.

For the pepper and chili pepper growing cycles, measured data have been used for calculating reference evapotranspiration (ET0) based on the Penman–Monteith equation according to the FAO procedure [21].

2.2.2. Plant Growth and Yield

On lettuce, when plants reached the commercial maturity stage (on 7 May 2024), non-destructive measurements of the SPAD index were performed (Spad 502, Minolta corporation, Ltd., Osaka, Japan) on 10 plants randomly selected for each experimental area. At harvest, plant height and marketable yield per plant were measured.

On pepper and chili pepper, two measurements were performed corresponding to two subsequent harvests (on 1 and 27 August 2024, and on 2 September and 11 October 2024, respectively). SPAD index, plant height, shoot fresh weight, and marketable yield per plant were measured.

2.2.3. QBS-ar Index and Soil Microarthropod Biodiversity

During the pepper and chili pepper growing cycles, soil samples were collected and categorized into three types of sampling points:

−

Control: Open-field areas without any panel coverage.

−

Area distant PV structure: Inter-row spaces that received partial shading.

−

Area close PV structure: Areas directly beneath the photovoltaic panels.

A total of 18 samples (3 samples × 2 crops × 3 areas) were collected and processed according to the procedure described hereafter. For each area and crop combination, three replicated samples (R1, R2, R3) were collected to ensure consistency and statistical robustness. At each sampling site, soil microarthropods were collected by extracting three individual soil cores positioned at the vertices of an equilateral triangle with sides measuring 10 m. Each core measured 10 cm × 10 cm × 10 cm, resulting in a volume of approximately 1000 cm³ per sample. The three cores, together amounting to a total volume of about 3000 cm³, were combined into a single polyethene bag for subsequent laboratory processing. The samples were transported to the laboratory within 48 h under controlled conditions, ensuring they remained undisturbed and protected from thermal and mechanical shocks. This allowed for the prompt initiation of the extraction procedure, minimizing the risk of alteration or degradation. The sampling design and QBS-ar methodology were applied according to standard procedures, following approaches used in recent studies on soil quality in Mediterranean environments [22].

Microarthropods were extracted from the soil samples using Berlese–Tullgren funnels over seven days under controlled laboratory conditions. The extraction method uses a thermal and desiccation gradient to drive soil-dwelling fauna downward into collection vessels. The soil samples were subjected to continuous heat exposure from a 60-watt incandescent bulb positioned 20 cm above each sample. The bulb remained switched on 24 h a day throughout the entire seven-day extraction period. A mesh with 2 × 2 mm openings was used during the process. The extracted specimens were collected and preserved in a solution of ethyl alcohol and glycerol at a 3:1 ratio. All samples analyzed in this study are currently stored at the University of Bari. Specimens were identified and counted using a stereomicroscope with magnifications ranging from 8× to 50×. Following extraction, specimens were sorted and identified to major taxonomic groups under a stereomicroscope. Identification was carried out to the level required for application of the QBS-ar index [14]. Each taxon was assigned an Eco-Morphological Index (EMI) score ranging from 1 to 20, reflecting the degree of morphological adaptation to soil life. Euedaphic taxa highly adapted to soil habitats (e.g., Protura, Diplura, Pauropoda) receive an EMI of 20, while less adapted or epigeic taxa receive lower scores [14]. The QBS-ar value for each sample was calculated as the sum of the EMI scores of all taxa present. This methodology is widely recognized for its simplicity, ecological significance, and robustness across diverse environments, including agricultural land, forests, urban soils, and wetlands [15,22]. The QBS-ar index provides an integrative measure of soil biological quality by combining faunal diversity and functional traits linked to soil ecosystem services.

2.2.4. Land Equivalent Ratio (LER) Calculation

The performance of the agrivoltaic system in terms of land surface requirements was assessed using the Land Equivalent Ratio (LER), computed as:

LER = [Yield agri (dual)/Yield agri (mono)] + [Yield elec (dual)/Yield elec (mono)]

Yeld agri (dual) and (mono) are the economic values of the agricultural production in the mixed and single production system, respectively; Yield elec (dual) and (mono) are the economic values of the energy production in the mixed and single production system, respectively.

For agricultural production values in the LER calculation, the yield values measured in the experiments of this study were used. In particular, two different annual agricultural programs were hypothesized, based on (i) a lettuce cycle and a pepper cycle, and (ii) a lettuce cycle and a chili pepper cycle. For the Yield agri (mono) the yield values obtained in the control conditions, the lettuce + pepper and lettuce + chili pepper were used. For the Yield agri (dual), the yield values obtained in the PV conditions were used (yield per cultivated surface), considering a 49.3% of the total area usable for cultivation, half of which is characterized by ‘Area close PV structure’ conditions and half by ‘Area distant PV structure’ conditions. The plant densities reported in the Section 2.1 were used for the calculations.

For the energy component, a 565,885 kWh ha⁻¹ yield has been considered, according to the PVsyst model based on the characteristics of the system [23]. The same value was used both for dual and mono Yield elec, since the system was designed as a pure photovoltaic system and only later converted to agrivoltaic use, without variation in electrical output.

All values used in the calculations were normalized for a 1 ha area.

2.2.5. Statistical Analysis

The crop measurements data were subjected to ANOVA, considering a completely randomized experimental design. Treatment means were separated by the Tukey HSD test when there was a significant effect at the p = 0.05 level. The statistical software STATISTICA 10.0 (StatSoft Inc., Tulsa, OK, USA) was used for the analysis.

3. Results and Discussion

3.1. Environmental Data/Microclimatic Variables

Based on measurements provided by the sensors network implemented in the experimental field, rather evident differences were detected among the different growing conditions tested in the study, confirming the influence of the photovoltaic system on the observed microclimatic variables.

In the lettuce growing cycle, cumulated radiation values of 365 MJ m⁻² and 306 MJ m⁻² were recorded in open field control and PV conditions, respectively (Figure 1A), with daily values remaining similar in the first half of the cycle up to the first decade of April, and then showing higher values in the Control, with an average difference of 25 MJ m⁻² up to the beginning of May and a gradual increase up to 58 MJ m⁻² at the end of the cycle in terms of cumulated radiation (Figure 1B). Much more evident differences were observed during the summer cycle. In this case, it was possible to monitor two sub-zones within the photovoltaic system, one closest to the row of panels exposed to sunlight, and therefore most affected by shading phenomena (‘Area close PV structure’), and one more distant, therefore less affected by shading (‘Area distant PV structure’) (see Section 2.1 and Section 2.2.1). In particular, the highest cumulated radiation was observed in the Control area (1616 and 2130 MJ m⁻² for pepper and chili pepper, respectively) and in the ‘Area distant PV structure’ treatment (1630 and 2044 MJ m⁻² for pepper and chili pepper, respectively), showing very similar trends (Figure 2A). In contrast, a drastically lower cumulative radiation value, approximately half, was observed in the area inside the photovoltaic system closest to the structures (883 and 1091 MJ m⁻² for pepper and chili pepper, respectively) (Figure 2A). Indeed, by observing the daily trend of solar radiation, it is possible to observe the substantially overlapping trends between the control and the cultivation area within the area more distant from the photovoltaic structures, while consistently lower values were observed in the area closest to the structures, with greater differences during the full summer period and gradually smaller as the autumn season approached (Figure 2B).

Our results are in agreement with numerous evidence that cultivation in photovoltaic environment involves a reduction in the availability of sunlight for the crop. This factor represents the main element of attention and concern regarding the effects of AV on crop performance [24]. However, our study clearly demonstrates that the shading effects can be highly variable depending on the position with respect to the photovoltaic structures. In the inter-row space, in fact, areas with a high level of shading and others with similar shade than open field conditions can be identified. Typically, the highest radiation reduction is recorded just below the panel’s array [25]; so, acting on the photovoltaic layout (density, inter-row spacing, etc.) has been identified as an optimization strategy to ameliorate crop growing conditions by limiting negative impact on energy production [26]. In this perspective, the use of models aimed at defining the effect of shading with spatial distribution analysis is deeply investigated. These models are considered an important tool for the identification of the most suitable agrivoltaic system layout for specific crops and geographical locations [27].

With regard to the other microclimatic variables monitored in the study, different trends were observed as affected both by the season and the position of the areas under comparison.

During the lettuce growing cycle, only data for daily mean air temperature and RH were available, showing no relevant differences for the two areas monitored (Control and PV, respectively) (

Table 1. Microclimatic variables during growing cycles of vegetable crops (lettuce, pepper, and chili pepper) cultivated in open field (Control) and under photovoltaic (PV) conditions. For pepper and chili pepper, two sub-zones within the PV system (‘Area close PV structure’ and ‘Area distant PV structure’) were monitored.

	Daily Mean T (°C)	Daily Mean RH (%)	Daily Mean Wind Speed (m·s−1)	Cumulated ET0 (mm)
	Lettuce growing cycle
Control	14.8	70.7
PV conditions	15.1	70.9
	Pepper growing cycle
Control	32.0	48.3	2.4	480
Area close PV structure	30.4	52.3	2.2	345
Area distant PV structure	30.4	52.6	1.6	449
	Chili pepper growing cycle
Control	29.6	56.7	2.3	597
Area close PV structure	27.4	64.9	1.8	397
Area distant PV structure	27.8	67.3	1.6	541

). However, during the pepper and chili pepper growing cycles, the differences among the areas under comparison were more evident. In particular, during the pepper growing cycle daily mean temperature resulted 1.6 °C higher in open field control conditions compared to the PV system areas, with no differences between the two sub-zones (close and distant PV structures); RH was lower in control conditions compared the PV conditions (48.3 vs. 52.45%, on average), with slightly higher values in Area distant PV structure compared to the other sub-zone; the highest daily mean wind speed was observed in control conditions, while the lowest in the sub-zone distant from PV structure (Table 1). Similar trends were observed in the chili pepper growing cycle, with higher values of daily mean temperature and wind speed and lower RH in the control area (Table 1).

The evapotranspirative demand was reduced, on average, in the PV growing conditions compared to open field, as an effect of reduced solar radiation and wind speed and increased air relative humidity (Table 1). Among the sub-zones within the PV system, a more relevant effect was observed in the sub-zone close to the photovoltaic structures, where cumulative ET0 was 28% and 34% lower than the Control with reference to the pepper and chili pepper growing cycle, respectively. However, in the sub-zone distant to the photovoltaic structures, the ET0 was 6% and 9% lower than the Control (Table 1). In general, the effects observed on microclimatic conditions are consistent with previous studies. Fagnano et al. (2024) [10] observed a reduced evapotranspiration (29%) in a ground-mounted photovoltaic plant where aromatic crops have been cultivated; in that case, the plant layout presented a 3.5 m inter-row spacing (compared to 8 m in our study), so the conditions can be considered similar to those of the sub-zone close to PV structures in our study, in which it was indeed observed a very similar ET0 reduction. Mechanisms of reduced ET in agrivoltaic conditions have been well described by Marrou et al. (2013) [28], who concluded that evapotranspiration reduction was mainly driven by the modifications of the micrometeorological conditions below the solar panels. In the scenario of climate change, resulting in global warming and water resource shortage trends, identifying strategies for improved water use efficiency and/or water saving is a critical issue for agriculture [29]. In this perspective, reduced ET as a consequence of microclimatic variations has been often outlined as a benefit of agrivoltaics [30]. However, it has been pointed out that reduction in ET under agrivoltaic conditions may not systematically lead to increased water use efficiency, depending on the genotypic plant sensibility of dry matter accumulation to shade [28]. As an effect of the capacity of agrivoltaic systems to reduce evapotranspirative demand under shade and to slow down soil water depletion, Ramos-Fuente et al. (2023) [31] reported improved irrigation water productivity for maize in AV, but only in fully irrigated compared to deficit conditions. In rainfed conditions, average maize grain yield was higher and more stable under AV than under full sun, with a proportional relation between yield increase and drought stress, indicating that AV systems could increase crop resilience to climate change [32].

3.2. Plant Growth and Yield

Lettuce growth was impaired by PV cultivation conditions. Plant height was reduced by 15%, and marketable shoot fresh weight per plant by 37%, on average, with no significant differences between the two sub-zones in the PV system (

Table 2. Plant growth and crop performance parameters of vegetable crops (lettuce, pepper, and chili pepper) cultivated in open field (Control) and under photovoltaic (PV) conditions in two sub-zones within the PV system (‘Area close PV structure’ and ‘Area distant PV structure’).

	Plant Height (cm)	Shoot Fresh Weight (g plant−1)	Fruit Fresh Weight (g plant−1)	SPAD
	Lettuce
Control	30.5 a	727.1 a		30.1
Area close PV structure	25.1 b	404.2 b		31.9
Area distant PV structure	26.6 b	518.6 b		30.3
Significance	***	***		ns
	Pepper
Date
1 August 2024	39.1	209.1	121.4	48
27 August 2024	61.2	372.7	242.2	49
Treatment
Control	52.6 a	374 a	248.6 a	51 a
Area close PV structure	56.8 a	295 ab	174.2 ab	45 b
Area distant PV structure	41.1 b	204 b	122.7 b	50 a
Significance
Date	***	***	***	ns
Treatment	***	***	***	**
Treatment × Date	***	ns	ns	ns
	Chili pepper
Date
2 September 2024	53.4	778.4	401.5	55
11 October 2024	55.4	862.6	409.0	53
Treatment
Control	63.6 a	1514.8 a	913.8 a	49.9 b
Area close PV structure	54.7 b	510.7 b	195.7 b	50.6 b
Area distant PV structure	45.4 c	472.8 b	131.9 b	60.7 a
Significance
Date	ns	ns	ns	ns
Treatment	***	***	***	***
Treatment × Date	ns	ns	ns	ns

Mean separation within columns by Tukey HSD test. Mean values followed by different letters within columns are significantly different (p < 0.05). Significance: ns = not significant, ** p ≤ 0.01, *** ≤ 0.001.

). Conversely, SPAD index was not influenced by the growing environments under comparison.

For pepper, highest values of plant height, shoot, and fruit fresh weight were observed at the second measurement date. In general, the best growing conditions were observed in open field control compared to PV, but with differences related to the PV sub-zone. In particular, the plant height was similar in the Control and in the Area close PV structure (54.7 cm, on average), while plants grown in the Area distant PV structure were more negatively affected by the modified growing conditions, showing the lowest shoot and fruit fresh weight, the latter reduced by 51% compared to the Control; intermediate values were observed for these parameters in the Area close PV structure, with a less severe tendency to yield (fruit fresh weight per plant) reduction. Conversely, SPAD resulted similar in Control and in Area distant PV structure, while values slightly lower were observed in the Area close PV structure.

For chili pepper, the highest plant height was observed in the Control, followed by plants grown in Area close PV structure, with lowest values in Area distant PV structure. Both shoot and fruit fresh weight were lower in PV conditions, regardless of the sub-zone, with yield reduction of 82% compared to the Control in terms of fruit fresh weight (Table 2).

Our results suggest that, in the conditions of our experiments, the microclimatic variations occurring under PV, compared to full-sun, open-field conditions, had relevant effects on plant growth. Although the reduction in solar radiation remains the key aspect, the interaction of other microclimatic conditions should be taken into account and further explored. Indeed, we observed reduced growth also on plants cultivated in the sub-zone less affected by shading, where solar irradiance values very similar to the full-sun control zone were recorded (Figure 2). We can hypothesize that reduced wind speed and increased RH under PV (Table 1) determined modified aerodynamic conditions, negatively affecting gas exchange and thermo-regulative mechanisms in plants. In addition, a different distribution of solar radiation during the day could have induced plant stress conditions in the cultivated area under PV and, in particular, in the area distant to structures. Further deepening should be implemented in order to explain the responses observed in the study, possibly including measurements related to gas exchange and photosynthetic activity parameters.

On the basis of the yield results reported in Table 1, taking into account the land available for cultivation in the PV system and in the Control, respectively, and the planting densities adopted for the different tested crops, the resulting yield per cultivated surface (on hectare basis) values are 14,230 and 45,444 kg ha⁻¹ for lettuce, in PV and Control conditions, respectively; 2288 and 7769 kg ha⁻¹ for pepper; 2525 and 28,556 kg ha⁻¹ for chili pepper. As a result, yield reduction of 68.7%, 70.5%, and 91.2% were observed in this study under PV conditions for lettuce, pepper, and chili pepper, respectively.

It has been reported that there is still little and contrasting information on the effects of combined PV energy and agricultural production on crop performance, and often the available information arises from studies with simulated or AV comparable conditions, such as agroforestry experiments or studies with artificial shade, or by the application of simulation and modeling approaches [6]. The current study was implemented with the aim of gathering crop performance information in real PV cultivation conditions, with specific reference to the system configuration described. The current experiments carried out on three crops (lettuce, pepper, and chili pepper) revealed a general reduction in plant growth and production parameters, at a different extent depending on the species. This is in agreement with the assumption that agrivoltaics may lead to a certain loss of production, in terms of yield both per plant and per cultivated surface, depending on the crop species and the photovoltaic system configuration [13]. With specific reference to the species tested in the current study, the average production reduction in terms of marketable yield per plant (mean value of the crop performance in the two sub-zones), observed under PV conditions can be considered in line with other studies. Marrou et al. (2013b) [33] reported reduced fresh shoot weight per plant for lettuce under agrivoltaic cultivation up to 42% and 19% in conditions of full or half density of panels configuration. Zheng et al. (2024) [11] reported lettuce yield reduction of approximately 50% under conventional agrivoltaic conditions, while no differences compared to full-sun conditions were observed when an improved agrivoltaic system with a grooved glass plate was used. For chili pepper, Joukhadar et al. (2025) [34] reported a marketable yield reduction of approximately 50% under AV conditions, with similar findings reported by Asa’a et al. (2024) [35], while an increased yield (more than 60%) was reported by Hernández et al. (2022) [36].

Generalizing the impact of agrivoltaic installations on crops is difficult. Cultivation in an AV environment may result in a different impact on the agronomic performance of winter compared to summer crops, or on short compared to long cycle crops [37]. It has also been noted that often yield response for the same crop may show variable and even opposite trends, depending on the typology of the agrivoltaic systems, the cropping cycle season, the specific annual climatic conditions during the study, the agronomical management practices, and even the variety used [1].

However, it is generally acknowledged that conventional agrivoltaic configurations based on traditional PV systems (as the case of fixed ground-mounted PV systems) generally lead to significant reduction in crop yield and quality compared to the Control where the climatic conditions and agronomic inputs availability do not represent limiting factors to crop performance. However, there is evidence that even conventional PV systems may lead to beneficial effects on crop yield in areas characterized by unsuitable conditions for plant growth such as excessive solar radiation, high temperature, drought, and water shortage [38]. The current attention towards innovative AV configurations and layout brings about lively interest in the perspective to ameliorate the crop performance under these systems.

3.3. QBS-ar Index and Soil Microarthropod Bioiversity

The mean QBS-ar values obtained across treatments and crops are reported in

Table 3. Summary of QBS-ar values and number of Taxa per crop and growing area per sample.

	QBS_R1	QBS_R2	QBS_R3	Mean QBS-ar	QBS-Max (R1 + R2 + R3)	Total Individuals
	Pepper growing cycle
Control	45	60	67	57.3	97	28
Area close PV structure	46	96	65	69	122	48
Area distant PV structure	60	70	82	70.7	102	34
	Chili pepper growing cycle
Control	46	45	45	45.3	56	25
Area close PV structure	51	75	30	52	145	30
Area distant PV structure	39	75	50	54.7	126	26

. The highest mean QBS-ar values were observed in ‘Area distant PV structure’—Pepper (QBS-ar = 70.7) and ‘Area close PV structure’—Pepper (QBS-ar = 69.0). Conversely, the lowest QBS-ar values were observed in Control—Chili Pepper (QBS-ar = 45.3) and ‘Area close PV structure’—Chili Pepper (QBS-ar = 52.0). These differences are clearly illustrated in Figure 3, where the bar chart shows QBS-ar trends across all treatments. Overall, both AV sub-zones exhibited higher QBS-ar values compared to Control for both crops under comparison. These findings suggest a positive influence of the PV system on soil biological quality. This trend is further supported by the maximum QBS-ar values (QBS-max) reported in Figure 3. The highest QBS-max values were recorded in “Area close to PV structure”—Chili Pepper (145), followed by “Area distant from PV structure”—Chili Pepper (126), and “Area close to PV structure”—Pepper (122). The lowest QBS-max value was found in Control—Chili Pepper (56). These results suggest that the microclimatic changes introduced by the PV structures—such as shading, reduced soil evaporation, and increased relative humidity—can create favorable conditions for soil fauna. Taxonomic richness (number of distinct taxa) across treatments and crops is summarized in Table 3. The highest richness was found in “Area close PV structure”—Pepper (48 taxa), while the lowest richness was observed in Control—Chili Pepper (25 taxa) followed by “Area distant PV structure”—Chili Pepper (26 taxa). These results further confirm a positive effect of photovoltaic installation combined with agricultural use of land on soil biodiversity, with marked differences between crops. The results demonstrate that agrivoltaic systems can positively influence soil biological quality. These findings align with previous research indicating that modified microclimatic conditions, such as shading and reduced soil evaporation, can create favorable environments for soil fauna. Dual use (agricultural and energy production) of land in PV installations, beyond its role in sustainable energy production, may also contribute positively to below-ground biodiversity. The QBS-ar index proved effective in capturing these effects, supporting its use as a monitoring tool in agroecological contexts.

The relative abundance of taxonomic groups revealed marked differences among the six sampling sites (Figure 4). The most numerically dominant taxa were Acari and Collembola, with peak abundances recorded in areas located close to PV structures. Sites adjacent to photovoltaic installations exhibited a higher overall abundance compared to both control and distant sites, with a significantly greater number of individuals across several taxa. These results suggest that the microenvironmental conditions influenced by the presence of PV panels (e.g., reduced evapotranspirative demand, shading, reduced thermal stress) may favor the development and persistence of soil arthropod communities. The graphical analysis supports the hypothesis of a positive indirect effect of photovoltaic structures on soil biodiversity.

Overall, the sites located near PV structures exhibited greater abundance and diversity compared to control or more distant sites. These results suggest that the microenvironmental conditions altered by the presence of the structures (e.g., shading, increased soil moisture, reduced thermal fluctuations) may favor the development and stability of soil communities.

3.4. LER

The LER, an index largely used to assess the effectiveness of any system that combines two (or more) production types on the same land unit (as the agrivoltaic case) [39], was used as an indicator of the productivity of the land used for the study. The LER indicator aims to compare the productivity of a mixed system versus a system based on a single production use.

When an annual cropping program consisting of a growing cycle of lettuce followed by pepper is hypothesized, the LER calculation results in the following:

$$\mathrm{L}\mathrm{E}\mathrm{R}=\left[{\displaystyle \frac{\mathrm{14,230} \mathrm{k}\mathrm{g} {\mathrm{h}\mathrm{a}}^{-1}}{\mathrm{45,444} \mathrm{k}\mathrm{g} {\mathrm{h}\mathrm{a}}^{-1}}}+{\displaystyle \frac{2288 \mathrm{k}\mathrm{g} {\mathrm{h}\mathrm{a}}^{-1}}{7769 \mathrm{k}\mathrm{g} {\mathrm{h}\mathrm{a}}^{-1}}}\right] +{\displaystyle \frac{565,885 \mathrm{k}\mathrm{W}\mathrm{h} \mathrm{h}{\mathrm{a}}^{-1}}{\mathrm{565,885} \mathrm{k}\mathrm{W}\mathrm{h} \mathrm{h}{\mathrm{a}}^{-1}}}=1.60$$

However, if the cropping program would consist of a lettuce followed by chili pepper growing cycle, the LER would result in the following:

$$\mathrm{L}\mathrm{E}\mathrm{R}=\left[{\displaystyle \frac{\mathrm{14,230} \mathrm{k}\mathrm{g} {\mathrm{h}\mathrm{a}}^{-1}}{\mathrm{45,444} \mathrm{k}\mathrm{g} {\mathrm{h}\mathrm{a}}^{-1}}}+{\displaystyle \frac{2525 \mathrm{k}\mathrm{g} {\mathrm{h}\mathrm{a}}^{-1}}{\mathrm{28,556} \mathrm{k}\mathrm{g} {\mathrm{h}\mathrm{a}}^{-1}}}\right] +{\displaystyle \frac{565,885 \mathrm{k}\mathrm{W}\mathrm{h} \mathrm{h}{\mathrm{a}}^{-1}}{\mathrm{565,885} \mathrm{k}\mathrm{W}\mathrm{h} \mathrm{h}{\mathrm{a}}^{-1}}}=1.40$$

LER > 1 indicates that considered dual system results in more efficient use of land compared to separate agricultural or energy production use. Based on several studies dealing with the assessment of AV competitiveness, the LER values of systems combining agricultural and photovoltaic energy production are impressive. Values ranging from 1.35 to 1.73 have been predicted in half (optimized for agricultural production) or full (optimized for energy production) density systems, respectively [39]. Since the effects of AV on crop yield differ depending on the type of crop, specific evaluation should be performed through subsequent cultivation, with the aim to find optimal cropping systems to maximize LER [40]. In this perspective, our study demonstrates that the LER can be greatly influenced by the choice of the crop, as occurred in our case as an effect of including pepper or chili pepper in the cropping program.

4. Conclusions

Although the assessment of advanced agrivoltaic systems represents the main focus and interest of scientific and industrial activity, the exploitation of the existing traditional photovoltaic plants in terms of combined agricultural and energy production should be considered a relevant challenge and opportunity, due to the large area currently occupied by these systems. Our study confirmed that the crop performance was negatively affected by the presence of the PV plant, with a variable response according to the crop species considered. However, also in conditions of yield reductions, the LER was improved, underlying a more efficient use of land, without negative or even ameliorative impacts on biological soil quality. Since the effects on plants depend on the typology of the agrivoltaic systems, the cropping cycle season, the specific annual climatic conditions, the agronomical management practices, and even the variety used, improving the knowledge on the specific response of crops, determined under real cultivation conditions typical of different environments, is necessary to define optimal cropping programs aimed to maximize resources use efficiency and land use.