Evaluation of Yield and Yield Components of Rice in Vertical Agro-Photovoltaic System in South Korea

Abstract

Renewable energy from photovoltaic power plants has increased in amount globally as an alternative energy to combat global climate change by reducing fossil fuel burning and carbon dioxide (CO₂) emissions. The agro-photovoltaic (APV) approach can be a solution to produce solar energy and crop production at the same time by installing solar panels on the same farmland to increase land use efficiency. This study aimed to compare the yield and yield components of rice (Oryza sativa L.) between a vertical APV system and a control field across two years. The solar panels were installed around the rice field in four directions of rice cultivation. Based on the analysis of variance, the primary factor influencing measured yield and yield components was the year effect, whereas the direction effect did not show significance, except for amylose content and ripened grains. Especially for rice production, the rice yield in 2023 was 6.8 t/ha, which was significantly higher by 0.8 t/ha than in 2022. Compared with the control condition, however, there was no significant negative impact on the year-to-year rice yield of the vertical APV system across two years. As rice yield was mainly affected by year, rice yield trials will be required for multiple years to understand the genetic and environmental factors influencing rice production under the vertical APV system.

1. Introduction

Food and energy are two of the several important requirements for human life. As the world population continues to increase toward the projected 10 billion by 2050, the production of global food and energy demand should increase accordingly [1]. Renewable energy from photovoltaic power plants has increased in amount globally as an alternative energy to combat climate change and global warming by reducing fossil fuel burning and CO₂ emissions [2,3,4,5]. In South Korea, the limited land area is a critical consideration for the installation of photovoltaic power plants. Photovoltaic power generation systems are being increasingly installed on agricultural land to increase land use efficiency. By installing solar panels in agricultural fields, both agricultural production and renewable energy generation via converting solar energy into electricity can coexist on the same land area [5,6,7,8,9,10]. This is regarded as agro-photovoltaic (APV) or solar sharing [5,6,7,8,9,10]. APV provides an opportunity to increase farm household income from crop production and electricity output compared with a traditional farming system [11]. The APV system may ensure food security and increase energy, as well as decrease greenhouse gas emissions. Thus, APV has been receiving attention from solar developers and farmers [12].

Generally, the APV system is installed above and around the crop. Rooftop photovoltaic panels for generating photovoltaic energy have mainly been constructed and installed at a height of approximately 3–4 m to reduce interference with farming activities [7,13]. However, the APV system generates shading conditions, which decreases the average available sunlight reaching the crops cultivated under the solar panels. Consequently, the amount of photosynthetically active radiation received by the plant for photosynthesis and plant growth is reduced. Light shortages or shading conditions are some of the considerations for crop cultivation and photosynthesis under the shading conditions of the APV system. The rate of photosynthesis is influenced by light intensity, CO₂, temperature, and humidity. However, the APV system has a positive impact on crop growth by reducing evapotranspiration, increasing carbon uptake, increasing water use efficiency, improving soil humidity, and decreasing temperature [6,8,14].

The beneficial effects of the APV system are dependent on the crop species. Barron-Gafford et al. [14] conducted an APV study with chiltepin pepper, jalapeño, and cherry tomato. Compared with the open-field condition, the APV system increased the CO₂ uptake and production of chiltepin pepper and cherry tomato and improved the water use efficiency of jalapeño and cherry tomato. In this regard, a considerable drop in crop yield is frequently experienced for lettuces under the APV system [15]. A study on rice paddies in Japan demonstrated a negative impact on the production and quality of rice with increasing shading rates [16]. Jo et al. [17] indicated that the biomass of rye and corn over two years was not statistically different between a rooftop APV system and open-field conditions, whereas rice yield was statistically reduced under the APV system compared with that under the open-field conditions.

Since the APV concept was first suggested in 1982 by Goetzerberger [18], various types of APV systems have been investigated to assess the responses of plants to light shortage and growing conditions under the APV system in comparison with the those under open-field cultivation. Several studies have reported on APV systems with monofacial panel arrays at full- and half-density. These studies indicated that crop yield for the half-density panel array was remarkably increased compared with that om the open field [11,18,19]. Katsikogiannis et al. [20] performed an experiment in which different topologies of APV array were installed under the climate of Boston, USA, to determine the overall land productivity. The results showed that the east/west (E/W)-facing and vertically installed arrays are most suitable for shade-intolerant or perennial crops because they permit more uniform shading conditions, whereas the south/north (S/N)-facing arrays were favorable for the cultivation of shade-tolerant crops.

Unlike monofacial solar panels, which collect solar irradiance from only the front side of the solar panel, bifacial panels are also capable of capturing light reflected off the ground onto the rear side of the panel [20,21]. Bifacial solar panels are considered the state of the art for electricity generation. It is convenient to install bifacial panels in the vertical APV system. In addition, bifacial modules are expected to become a more efficient, cost-effective approach for collecting renewable energy; their use is projected to increase to around 55% by 2031 [22].

Rice (Oryza sativa L.) is one of the most widely cultivated crops and is a major staple food crop, especially in Asian countries, including South Korea. The cultivation area for rice in South Korea in 2023 was 708,012 ha, which was 2.6% higher than that in 2022 [23]. Installing solar panels in rice paddies shows great potential for both rice production and electricity production. However, crop yield under the APV system is highly dependent on the local environmental conditions and crop species. In addition, not enough studies on the vertical APV system have been carried out with different crop species in South Korea. This study aimed to compare rice yield and its yield components between the vertical APV system and the control conditions across two years. The vertical APV system with solar panels was installed around the rice paddy field (a rectangular area of a rice paddy field). The split-plot design was used for this study. The solar panel’s directions of rice cultivation were applied to the main plot, and sub-blocks of each direction were used as the sub-plots of split-plot design. This is the first study to evaluate rice yield, seed compositions, and rice yield components under a vertical APV system with bifacial solar panels in a rice paddy field.

2. Materials and Methods

2.1. Solar Panel Installations

The vertical APV system was installed around the rice paddy field at the affiliated field of Kyungpook National University, Gunwi, South Korea (36°14′05″ N, 128°78′14″ E) (Figure 1a,b). A vertical APV system with 288 bifacial solar panels in total was installed in a rectangular area of a rice paddy field covering 3000 m² (30 × 100 m) with a height of 2.957 m (Figure 1c,d). Based on the rice cultivation, the solar panels were installed in four different directions of rice cultivation: southwest (SW), southeast (SE), northeast (NE), and northwest (NW). The SW and NE arrays face the SW/NE orientation, and the SE and NW arrays face the SE/NW orientation (Figure 1e).

2.2. Field Experimental Design with Rice (O. sativa L.) between the APV System and the Control Conditions

The experimental design was a split-plot design with replications. The solar panel’s directions of rice cultivation were applied to the main plot, and sub-blocks of each direction were used as the sub-plot of the split-plot design. To evaluate the effects of the vertical APV system on yield, amylose, protein contents, and yield components, the rice (O. sativa L.) cultivar ‘Wonkwang’ was grown in APV system conditions and open-field conditions (control). All rice seeds were planted on 23 April 2022 and 22 April 2023, Transplantation of rice seedlings into the field was conducted on 3 June 2022 and 5 June 2023. The planting density was 30 × 15 cm, with around 15 plants per hill. The rice grains were harvested on November 1, 2022, and October 18, 2023. The fertilizer was applied to fields with N-P₂O₅-K₂O fertilization in an amount of 90-45-57 kg ha⁻¹. Herbicide and insecticide were applied to control the weed and insect. The climate data during the rice growing season across two years are listed in Table S1.

Based on the observations for 8 a.m. and 5 p.m. during the rice growing season in 2022 (from 21 June to 21 September), the longest length of shade in the solar panel directions was different. The longest shade length was 7 m in the SW on 21 September, 5.5 m in the SE on 21 August, 5.5 m in the NE on 21 August, and 4 m in the NW on 21 August. Because the longest length of shade in each solar panel’s direction was different, different sampling methods were used. The numbers of sub-blocks sampled from the solar panels were 10 sub-blocks in the SW direction, 5 sub-blocks in the SE direction, 7 sub-blocks in the NE direction, and 5 sub-blocks in the NW direction. Each sub-block consisted of 2 rows. The sampled area was also applied differently depending on each solar panel’s directions. The SW direction was 9.0 × 0.6 m, the SE was direction was 5.3 × 0.6 m, the NE direction was 9.0 × 0.6 m, and the NW direction was 5.3 × 0.6 m. Based on the solar panel directions, the harvested areas were 5.4 m² for the SW/NE orientation and 3.18 m² for the SE/NW orientation in 2022 and 2023. There were three replications, two replications, three replications, and two replications for the SW, SE, NE, and NW solar panel directions, respectively.

2.3. The Open Filed Condition as Control

The central area of the rectangle APV system, where there was no shading condition for the solar panels, was regarded as the control (open-field condition); 5-sub-block samples were taken with 2 replications.

2.4. The Measurements of Yield and Yield Components of Rice

When the rice was matured under the vertical APV system and control conditions, ten randomly selected hills from each sub-block were assessed for plant height, culm length, panicle length, number of tillers per hill, yield, rice head rate, protein content, moisture content, amylose content, and percentage of ripened grain. The ripened grain (%) was the number of ripened grains per hill divided by the total number of spikelets per hill. The amylose and protein content of rice grains were determined using Grain Analyzer 1241 (Foss Analytical, Höganäs, Sweden), and the head rice rate was evaluated with Cervitec Grain Inspector TM 1625 (Foss Analytical, Höganäs, Sweden).

2.5. Statistical Analysis

In this study, all phenotypic measurements were statistically analyzed using SAS v9.4 (SAS Institute, Cary, NC, USA, 2013). The analysis of variance (ANOVA) was analyzed via PROC ANOVA in SAS. Combined ANOVA was then computed across years using the adjusted means. In the combined analysis, years were treated as random effects, while main plots (directions of the solar panel) and sub plots (sub-blocks of each solar panel direction with 2 rows) were considered fixed effects. The least significant difference test (LSD) was used for the differences among the mean values of measured phenotypic measurements at a 5% level of significance.

3. Results

In order to understand the vertical APV system, the present study compared rice yield, yield components, and protein content between the vertical APV system and the control conditions over two years. ANOVA was conducted to identify the decrease in yield and yield components of rice because of the shading conditions of the vertical APV system (

Table 1. Mean squares from combined analysis of variance of yield and yield components of rice between the control and the vertical APV system over two years.

Source of Variation	Degree of Freedom	Plant Height	Culm Length	Panicle Length	Number of Tillers	Yield	Head Rice	Moisture	Amylose	Protein	Ripened Grains
Year	1	507.0 ***	1009.5 ***	88.1 ***	205.7 ***	24.3 ***	4180.8 ***	0.25 *	914.9 ***	4.14 ***	1455.3 ***
Rep(year)	4	3.9 ns	10.7 *	9.1 ***	9.6 **	2.3 ***	20.9 ns	0.27 ***	3.2 *	0.23 ***	32.0 *
Direction (d)	4	48.2 ns	43.9 ns	3.8 ns	7.1 ns	7.7 ns	90.4 ns	0.03 ns	5.5 **	0.32 ns	52.9 **
Year × d	4	31.8 ***	33.3 ***	7.7 ***	4.6 ns	1.3 **	131.8 ***	0.18 **	8.0 ***	0.09 ns	2.3 ns
Sub-block (sb)	9	8.5 *	4.9 ns	1.7 ns	5.8 ns	0.7 ns	34.6 ns	0.03 ns	0.8 ns	0.19 *	5.4 ns
Year × sb	9	2.4 ns	2.5 ns	0.9 ns	3.7 ns	0.8 *	29.3 ns	0.02 ns	1.0 ns	0.06 ns	3.2 ns
d × sb	18	10.1 ns	6.5 ns	1.1 ns	5.6 ns	0.7 ns	17.9 **	0.06 ns	1.0 ns	0.12 **	7.9 ns
Year × d × sb	18	7.6 *	5.6 ns	0.9 ns	6.3 **	0.9 ***	5.7 ns	0.03 ns	0.6 ns	0.03 ns	9.6 ns
Error	104	4.3	3.6	1.4	2.4	0.3	24.1	0.04	1.3	0.04	9.5

ns, not significant. * Significant at the 0.05 probability level. ** Significant at the 0.01 probability level. *** Significant at the 0.001 probability level.

). The year effect was significant for all measured yield and yield components (p < 0.05). In this study, vertical solar panels were installed in four different directions: SW, SE, NE, and NW (Figure 1e). Based on ANOVA, the direction effect did not show significance for all measured yield and yield components except amylose content (p < 0.05) and ripened grains per hill (p < 0.05). The sub-block effect showed significance for plant height (p < 0.05) and protein content (p < 0.05). For the interaction effects, year × panel direction was significant for plant height (p < 0.001), culm length (p < 0.001), panicle length (p < 0.001), yield (p < 0.01), head rice (p < 0.001), moisture (p < 0.01), and amylose content (p < 0.001); year × sub-block was significant only for yield (p < 0.05); panel direction × sub-block showed significance for head rice (p < 0.01) and protein (p < 0.01); year × panel direction × sub-block was significant for plant height (p < 0.05), number of tillers (p < 0.01), and yield (p < 0.001).

Based on the ANOVA results, the primary factor influencing all measured yield and yield components was the year effect in this study (Table 1). A larger mean square value in ANOVA means a greater factor effect on the measured yield and yield components. For the year effect, all measured yield and yield components under the vertical APV system were significantly different between two years (Figure 2). Specifically, the plant height (96.1 cm), culm length (80.7 cm), number of tillers (17.8), head rice (82.3%), moisture (10.44%), and amylose content (19.9%) in 2022 were statistically higher than those in 2023. The yield and protein content in 2023 was 6.8 t/ha and 5.7%, respectively, which were significantly higher by 0.8 t/ha and 0.3%, respectively, than those in 2022.

Combined ANOVA for a split-plot design over two years demonstrated that the solar panel direction effect as the main plot of the split-plot design showed significance for amylose content and the percentage of ripened grains per hill in Table 1. However, for individual years, ANOVA showed that the direction effect was the most important factor impacting the rice yield and rice yield components in the vertical APV system, and most measured yield and yield components showed significance, except head rice and moisture in 2022, and number of tillers in 2023 (Table S2). Thus, the mean values of rice yield and its yield component grown under the solar panels installed in four different directions (NE, SW, SE, NW) are shown in Figure 3. The mean plant height ranged from 92.2 cm (control) to 95.4 cm (NE), with no significant differences among the panel directions. The highest culm length and panicle length were observed for the NE solar panel (79.4 cm) and NW solar panel (16.9 cm), respectively. There were no significant differences among all four directions of panels for the number of tillers, which ranged from 15.8 (SE) to 17.1 (SW). The highest rice yield was in the control conditions (6.8 t/ha), and was not significantly different from that obtained under the SW solar panel (6.4 t/ha), SE solar panel (6.2 t/ha), and NE solar panel (6.7 t/ha). For moisture and amylose content, they did not show significance compared with the control conditions. Protein content in rice grown under the different panel directions ranged from 5.39% to 5.62% across two years. In terms of panel direction, the SW direction led to significantly higher protein content compared with the control conditions.

The most important factor is the year effect for the rice yield and its yield components in this study, followed by the panel direction and the year × panel direction interaction effects (Table 1). The mean values of the rice yield and rice yield components by year and panel direction are listed in

Table 2. Yield and yield components of rice between the agro-photovoltaic system and the control (open field) in four different directions of solar panels from each year.

Environments	Plant Height (cm)	Culm Length (cm)	Panicle Length (cm)	Number of Tillers	Yield (t/ha)	Head Rice (%)	Moisture (%)	Amylose (%)	Protein (%)	Percentage of Ripened Grains per Hill (%)
2022 southwest	95.4 ± 4.3a	78.8 ± 4.0a	16.5 ± 2.2a	16.7 ± 2.4ab	6.4 ± 0.8a	78.5 ± 6.7a	10.38 ± 0.24ab	17.8 ± 2.5a	5.7 ± 0.4a	87.3 ± 4.9bc
2022 southeast	94.9 ± 3.4ab	78.6 ± 3.9ab	16.3 ± 1.3ab	16.3 ± 2.6b	6.4 ± 0.6a	76.2 ± 7.4a	10.46 ± 0.24a	17.2 ± 3.6a	5.6 ± 0.2ab	88.0 ± 3.7ab
2022 northeast	94.7 ± 2.8ab	78.5 ± 3.4ab	16.2 ± 1.4ab	16.0 ± 1.7b	6.2 ± 1.0a	77.5 ± 7.5a	10.37 ± 0.20ab	17.7 ± 2.5a	5.6 ± 0.3ab	87.6 ± 4.5abc
2022 northwest	94.2 ± 2.8abc	78.3 ± 3.5ab	15.8 ± 1.5ab	16.7 ± 2.0ab	6.3 ± 1.1a	76.0 ± 8.0a	10.44 ± 0.21ab	17.6 ± 2.6a	5.5 ± 0.2ab	87.4 ± 3.8abc
2022 control	93.6 ± 3.1bc	77.3 ± 3.6b	16.3 ± 1.3a	16.7 ± 1.9ab	6.2 ± 1.1a	76.6 ± 8.2a	10.41 ± 0.23ab	17.5 ± 2.5a	5.5 ± 0.2ab	88.1 ± 4.2ab
2023 southwest	94.2 ± 2.7abc	78.1 ± 3.0ab	16.2 ± 1.1ab	17.4 ± 1.5ab	6.8 ± 0.8a	76.0 ± 4.8a	10.36 ± 0.19ab	17.4 ± 2.5a	5.4 ± 0.2ab	87.0 ± 5.2bc
2023 southeast	93.0 ± 2.1c	77.6 ± 2.5ab	15.4 ± 1.3b	17.3 ± 2.4ab	6.3 ± 10.7a	80.4 ± 6.2a	10.33 ± 0.18b	17.7 ± 2.5a	5.4 ± 0.2ab	87.2 ± 4.7bc
2023 northeast	94.5 ± 3.3ab	78.4 ± 3.0ab	16.2 ± 1.5ab	18.3 ± 3.1a	6.6 ± 0.8a	79.4 ± 3.6a	10.38 ± 0.22ab	17.4 ± 1.7a	5.6 ± 0.2ab	87.9 ± 4.8ab
2023 northwest	94.2 ± 2.5abc	78.3 ± 2.9ab	16.0 ± 0.9ab	17.8 ± 2.3ab	6.7 ± 0.7a	78.9 ± 6.4a	10.35 ± 0.23ab	17.8 ± 2.4a	5.2 ± 0.9c	86.1 ± 6.4c
2023 control	93.8 ± 2.5bc	77.7 ± 2.8ab	16.1 ± 0.9ab	17.5 ± 2.6ab	6.8 ± 0.8a	78.4 ± 6.7a	10.42 ± 0.20ab	18.0 ± 2.4a	5.5 ± 0.2ab	89.1 ± 3.5a
LSD	1.5	1.5	0.9	1.8	0.8	5.1	0.13	0.9	0.2	1.7

LSD, least significant difference. The same letter within measured traits indicates no significant difference at p ≤ 0.05.

. In 2022, the SW panel resulted in the tallest plant height, the tallest culm length, and the tallest panicle length, which averaged 95.4, 78.8, and 16.5 cm, respectively. In 2023, the NE solar panel resulted in the highest number of tillers per hill (18.3), which was significantly higher than that obtained under the SE and NE solar panels in 2022. Rice growth was comparable between the vertical APV system and the control conditions. The rice yield ranged from 6.2 to 6.8 t/ha. The highest and the lowest rice yields were in the control conditions in 2023 and 2022, respectively. Rice yield was 2.5–4.3% higher in the APV system compared with that in the control conditions in 2022, and the highest yield reduction of 7.2% was observed in the APV system compared with that in the control conditions in 2023, albeit without a significant difference. For head rice and amylose content, the year × panel direction effect was not significant. The SE solar panel in 2022 produced the highest moisture content, which was significantly higher than that in 2023 for the SE solar panel. For the protein content, the SW solar panel in 2022 produced the highest moisture content, which was significantly higher than that in 2023 for the NW solar panel. The percentage of ripened grains per hill ranged from 86.1% to 89.1%. In 2022, the percentages of ripened grains per hill were not significantly different among the four different panel directions. However, the amount of ripened grains per hill in 2023 for the NW solar panel was significantly lower than that in 2023 for the NE solar panel and the 2023 control conditions.

According to the simulation with sunrise and sunset, a different number of sub-blocks in each direction was determined. As the year × direction × sub-block interactions were significant for plant height, number of tillers, and yield, the mean values of plant height, number of tillers per hill, and yield of rice are listed in

Table 3. Yield, yield components, and yield index of rice between the agro-photovoltaic system and the control (open field) in sub-blocks from solar panel directions between 2022 and 2023.

Direction	Sub-Block	2022				2023
		Plant Height (cm)	Number of Tillers	Yield (t/ha)	Yield Index (%)	Plant Height (cm)	Number of Tillers	Yield (t/ha)	Yield Index (%)
Southwest	1	103.2 ± 8.0	18.8 ± 0.4	6.4 ± 0.4	103.9	93.6 ± 1.9	15.4 ± 1.9	6.6 ± 0.9	97.4
	2	99.4 ± 1.9	17.2 ± 2.5	6.2 ± 0.2	100.9	92.6 ± 2.3	14.7 ± 1.9	6.7 ± 0.7	97.7
	3	98.5 ± 0.8	16.7 ± 0.4	6.7 ± 0.9	108.9	93.5 ± 2.2	15.6 ± 1.2	6.6 ± 0.6	96.8
	4	97.0 ± 1.4	16.9 ± 1.1	5.1 ± 0.7	81.9	92.7 ± 1.6	15.7 ± 1.6	7.0 ± 0.7	102.1
	5	97.1 ± 0.6	19.3 ± 1.9	6.0 ± 0.7	96.9	91.5 ± 2.2	15.5 ± 1.2	6.5 ± 0.8	94.7
	6	95.7 ± 1.0	19.1 ± 1.3	5.7 ± 0.3	92.9	92.7 ± 2.9	16.2 ± 0.4	7.2 ± 0.5	105.0
	7	94.9 ± 0.1	20.5 ± 1.7	5.3 ± 0.4	85.9	91.7 ± 2.1	16.8 ± 1.2	7.2 ± 0.6	105.0
	8	96.3 ± 1.3	20.3 ± 2.7	6.0 ± 0.2	97.9	92.7 ± 4.1	16.2 ± 1.9	7.2 ± 0.8	105.0
	9	95.7 ± 0.8	19.6 ± 1.0	6.7 ± 0.9	108.9	92.8 ± 2.9	15.9 ± 1.4	6.7 ± 0.5	98.6
	10	95.2 ± 2.1	18.1 ± 3.5	6.3 ± 0.4	101.9	92.3 ± 2.2	16.9 ± 1.7	7.3 ± 0.7	107.7
	LSD0.05	4.4	3.4	10.3		4.2	3.3	1.2
Southeast	1	94.3 ± 1.7	16.8 ± 1.2	6.8 ± 0.7	109.4	95.5 ± 3.6	14.7 ± 1.8	6.8 ± 0.2	99.1
	2	97.2 ± 0.1	17.4 ± 1.1	6.0 ± 0.4	96.7	94.7 ± 3.9	13.3 ± 0.6	6.4 ± 1.7	93.4
	3	93.4 ± 0.5	15.3 ± 0.1	5.0 ± 0.9	81.5	96.0 ± 2.6	15.7 ± 0.8	6.8 ± 0.7	99.1
	4	92.2 ± 0.8	17.0 ± 1.7	5.8 ± 0.7	94.2	92.5 ± 0.4	14.7 ± 1.2	7.3 ± 0.3	107.2
	5	92.4 ± 3.7	17.2 ± 2.1	5.5 ± 0.2	89.1	93.8 ± 2.9	16.4 ± 1.1	5.3 ± 1.3	78.4
	LSD0.05	4.8	3.7	1.6		7.6	3.1	2.6
Northeast	1	97.5 ± 1.7	17.1 ± 1.2	6.5 ± 0.2	105.9	95.1 ± 1.2	15.7 ± 1.2	7.1 ± 0.4	103.6
	2	98.6 ± 0.6	17.5 ± 1.5	6.7 ± 0.3	107.9	92.6 ± 2.5	15.3 ± 2.3	6.8 ± 0.5	99.5
	3	98.9 ± 1.1	18.2 ± 2.6	6.7 ± 0.5	107.9	93.7 ± 0.9	14.7 ± 0.4	6.6 ± 0.7	96.4
	4	99.2 ± 1.4	17.8 ± 0.8	6.4 ± 1.1	102.9	94.3 ± 1.6	16.4 ± 0.8	7.0 ± 0.1	102.3
	5	97.1 ± 2.0	17.2 ± 0.5	6.3 ± 1.1	101.9	92.7 ± 1.5	15.5 ± 0.6	7.3 ± 0.3	107.2
	6	95.7 ± 2.1	18.2 ± 0.3	7.2 ± 0.8	116.9	93.9 ± 1.7	16.0 ± 0.3	7.5 ± 0.5	109.5
	7	93.8 ± 0.8	14.9 ± 1.0	5.7 ± 0.9	91.9	92.7 ± 1.2	16.6 ± 0.6	6.9 ± 0.2	100.5
	LSD0.05	2.6	2.3	1.4		2.8	1.9	75.7
Northwest	1	97.1 ± 1.0	22.2 ± 0.3	5.5 ± 0.2	89.1	92.5 ± 0.9	15.7 ± 1.1	5.7 ± 0.4	83.0
	2	98.6 ± 0.6	20.9 ± 5.6	5.5 ± 0.2	89.1	94.2 ± 1.7	16.4 ± 1.1	6.7 ± 0.6	98.0
	3	94.7 ± 1.8	13.9 ± 0.9	3.9 ± 0.2	63.6	94.0 ± 1.1	16.0 ± 1.3	6.3 ± 0.7	92.2
	4	95.2 ± 0.1	15.4 ± 1.0	4.2 ± 0.2	68.7	92.0 ± 0.3	15.1 ± 2.3	6.2 ± 10.0	91.1
	5	94.7 ± 0.3	15.5 ± 1.1	3.5 ± 0.4	56.0	92.8 ± 0.1	16.4 ± 0.1	6.3 ± 0.4	92.2
	LSD0.05	2.5	6.7	0.7		2.6	3.5	1.7
Control	1	92.7 ± 0.9	15.9 ± 1.0	5.8 ± 0.3	100.0	92.0 ± 1.1	16.6 ± 1.9	7.4 ± 0.7	100.0
	2	92.7 ± 0.8	16.9 ± 0.8	6.4 ± 0.5		90.8 ± 0.9	15.9 ± 1.4	6.5 ± 0.5
	3	93.1 ± 1.7	17.0 ± 2.1	6.2 ± 0.7		90.5 ± 1.6	15.8 ± 0.9	7.0 ± 0.2
	4	93.5 ± 2.7	20.4 ± 1.6	6.5 ± 0.1		91.0 ± 1.7	16.1 ± 1.3	7.7 ± 0.8
	5	96.1 ± 2.2	19.2 ± 1.1	7.3 ± 0.3		90.1 ± 1.2	15.0 ± 0.5	6.6 ± 0.3
	LSD0.05	3.3	2.6	0.7		2.4	2.4	10.1

LSD, least significant difference at p ≤ 0.05.

. The mean values of measured phenotypes in Table 3 revealed the variability in sub-blocks within solar panel directions. The yield indexes in 2022 showed variability and ranged from 56.0% to 116.9%. More than 20% yield reductions were seen in the NW direction in 2022. However, yield reductions in 2023 were less than 10%, except for the first sub-block (83.0%) from the NW solar panel. The highest yield index across two years was 116.9% in the sixth sub-block from the NE solar panel in 2022, whereas the lowest yield index was 56.0% in the fifth sub-block from the NW solar panel in 2022.

4. Discussion

Based on the Renewable Energy 3020 Implementation Plan and the 2050 Carbon Neutral Strategy, the South Korean government has supported APV implementation since 2018 and has announced that by 2030, the proportion of renewable energy will increase to 20% and that energy from photovoltaic power generation will increase to 30.8 GWp [24,25]. However, South Korea has limited arable land. To increase land use efficiency, installing an APV system on farmland can be a solution to producing crops and renewable energy simultaneously. Thus, the Korean government has been interested in the expansion of photovoltaic power plants, including the APV system.

This study compared the yield and yield components of rice between the vertical APV system and control conditions over two years. Crop yield is of major importance in crop breeding programs, and yield components indirectly or directly impact crop productivity. Light reduction by an APV system, including a vertical APV system, affects crop growth. According to some previous studies, the shading conditions of the APV system had a negative effect on rice yield and yield components compared with the control conditions [16,26]. Our previous study evaluated agronomic performances and crop yields under the rooftop APV system in a paddy field and a dry field for various crops under the climate of Gunwi, South Korea [17]. The results showed that rice yield differences between the control and under the rooftop APV system were 1.4 t/ha in 2018 and 0.7 t/ha in 2019, which corresponded to yield reductions of 18.7% and 8.9%, respectively, under the rooftop APV system. There are other studies supporting a rice yield reduction under the APV system compared with that under control conditions [26,27,28]. In our present study, we installed a vertical APV system around a rice paddy field. Ahmed et al. [29] reported that rice production under the vertical APV system was higher than that in a tilted and horizontal APV system. However, they indicated, with three different APV systems, that the lowest energy output was from the vertical APV system. As a result, rice yields under solar panels installed in four different directions were even higher than those in the control condition in 2022, whereas the highest yield reduction was 7.2% in the SE direction of solar panels in 2023, although our previous study and this study used the same rice cultivar and same growing location (Table 2). The targeted rice yield under the APV system should be more than 80.0% of the crop yield in normal growth conditions. Taken together, it was possible to produce more than 80% of the rice yield in our previous study with the rooftop APV system and of the rice field in this study with a vertical APV system in South Korea.

In our previous study, plant height was not significantly different between the open field and the rooftop APV system in 2018, whereas it did show significance in 2019 [17]. Lee et al. [28] installed rooftop APV facilities in three different regions of South Korea that resulted in different rice plants height between the APV system and the control at two locations, whereas the other location of the APV system produced a higher plant height than that in the control plot. The percentage of ripened grains ranged from 90.9% to 93.8% across two years [17]. Another study showed that there were no significant differences between the APV system and the open-field conditions for experiments carried out on farms at four different sites with different rice cultivars for the percentage of ripened grains of rice [16]. Under the vertical APV system used in the present study, the maximal percentage of ripened grains was less than 90.0% (Table 2 and Table 3, Figure 2 and Figure 3). For seed quality, protein content was evaluated in this study. Higher protein content in rice seed has been shown to negatively impact texture and taste due to increased viscosity [30,31]. Gonocruz et al. [16] showed that shading due to the APV system in four different farms in Japan increased the protein content in rice compared with that in open-field conditions. In the present study, different directions of solar panels did not affect the accumulation of protein content in rice compared in the open-field conditions, whereas sub-block and panel direction × sub-block effect for protein content showed significance based on ANOVA (Table 1).

High air temperatures have a negative impact on plant growth [32]. In general, the APV system retains soil moisture, and the air and soil temperatures are lower under the APV system compared with those under open-field conditions. The APV system is reported to have a positive impact on crop growth by reducing evapotranspiration, increasing carbon uptake, and increasing water use efficiency [5,6,7,8,9,10]. The vertical APV system may have similar beneficial effects. However, Heidari et al. [33] reported that snow accumulation on solar panels decreased the generation of electricity, and the annual energy losses due to snow accumulation ranged from 5% to 34%. The reflectivity of snow and the ground increases the amount of reflected radiation reaching the solar panels of a vertical APV system. In addition, dust on solar panels may decrease electricity output. The advantage of the vertical APV system is that it requires less cleaning of solar panels compared with tilted panels.

Under the impact of an APV system, the current study estimated the yield components of a main crop, rice. Our findings may also provide additional quantitative data to generate a decision support system to implement rice growth under a vertical APV system in order to improve energy output without compromising crop performance. However, determining suitable crops with a consideration of shading compared with that under control conditions is crucial to understanding rice performance, which affects crop yields in plant environmental adaptation. This study did not find a negative impact of the vertical APV system on grain yield and yield-related characteristics in rice. Crop yield is determined by a combination of genetic and environmental factors. The limitation of this study was in its utilization of a single rice genotype under the vertical APV system, and that it was difficult to determine the effect of genotype on yield and yield components of rice. In addition, combined ANOVA resulted in the year effect being the most important factor affecting rice yield in this study. For further research, crop yield trials will still be required for multiple years and rice cultivars in order to understand the influence of genetic and environmental factors on rice yield under the vertical APV system.

5. Conclusions

In conclusion, rice was evaluated to understand the yield and yield components in the vertical APV system. In this study, the results demonstrated that there was no statistical difference for the rice yield between the vertical APV system and the control conditions across two years. The target rice yield with the APV system should be more than 80.0% of the crop production under normal growth conditions. This study showed that the highest yield reduction was 7.2% under the SE solar panel in 2023. Thus, it is possible to produce more than 80% of rice yield with a vertical APV system in South Korea. However, ANOVA revealed that the primary factor influencing all measured yields and yield components was the year effect. The rice yield in 2023 was 6.8 t/ha, which was significantly higher by 0.8 t/ha than that in 2022. As rice yield was mainly affected by year, rice yield trials will be required for multiple years to understand the genetic and environmental factors influencing rice production under the vertical APV system.