Spatial and Temporal Variability of Environmental Variables in Chinese Solar Greenhouses in the Summer Season

Abstract

To ensure a high crop profit in Chinese solar greenhouses (CSGs), it is crucial to effectively manage major environmental variables such as temperature, humidity, and CO₂ concentrations, among others, to mitigate harmful effects on crop growth. The objectives of this study were to assess the spatial, vertical, and temporal variability of major environmental variables in CSGs during summer, and to provide fundamental information that could facilitate the monitoring and control of environmental factors in CSGs. The experiments were conducted in two CSGs: one with crops and another without crops. The measured environmental variables included air temperature, humidity, CO₂ concentration, light intensity, and wind conditions. Significant variations in the spatial, vertical, and temporal distribution of environmental factors were observed in both greenhouses. The results revealed significant diurnal patterns in temperature and humidity, with higher daytime temperatures and lower humidity levels. The greenhouse with crops exhibited warmer bottom layers due to restricted air mobility. CO₂ concentrations peaked at night, aligning with plants’ respiration and photosynthesis cycles, whereas light intensity showed substantial daytime peaks, slightly affected by the presence of crops. The study emphasized the necessity of stratified control of the environment and dynamic management of CO₂. The deployment of a wireless sensor network (WSN) and placement of an error-based sensor ensured precise monitoring, highlighting the importance of continuous data collection and adaptive management for optimal greenhouse conditions.

1. Introduction

Implementing the techniques of precision agriculture is being increasingly recognized as a vital strategy for mitigating climate-related impacts and fostering environmentally responsible farming practices. In the context of Chinese solar greenhouses (CSGs), such methods are proving instrumental in enhancing productivity and operational efficacy, especially within the climatically challenging regions of northern China. These greenhouses apply strategic advancements to increase crop outputs significantly, demonstrating their critical role in advancing ecologically sound agricultural practices [1]. CSGs are controlled environments that increase crop yield, especially high-priced and off-season vegetables in northern China. During the day, northern China receives sunlight more than 50% of the time. However, the temperatures are very low in winter, with daily average temperatures in the three coldest months dropping below −25 °C [2].

The microclimate within CSGs experiences significant variations that are largely dependent on the external climatic conditions, especially when no auxiliary heating and humidity systems are provided [3]. The widespread popularity of CSGs in China is significantly attributed to their energy-efficient characteristics, which contribute to a cost-effective operation [4,5]. The need to increase productivity is driving the construction of higher-quality structures, which require a longer time to recoup their financial investments. A challenge arises when growing protected crops during the summer, as solar radiation heats the crop inside the structure and hinders the exchange of air with the cooler outside atmosphere [6]. The microclimate in the CSG environment, particularly the temperature, humidity, and radiation, is crucial for ensuring optimal crop growth and high yields with quality [7]. The distribution of temperatures throughout the greenhouse and controlling the temperature and humidity within an appropriate range are essential for ensuring the uniform growth of crops [8].

The energy-saving properties of CSGs significantly reduce operational costs. However, temperature control and adequate ventilation pose significant challenges during the summer, when solar radiation heats the interior and the covering impedes exchange with the cooler outside air. Several studies have characterized the environmental conditions in greenhouses. Ryu et al. [9] monitored changes in the air temperature, humidity, photosynthetic photon flux density, and CO₂ in greenhouses and found that the inside temperature increased in the daytime. When the vents were open during the day, the minimum and maximum temperature and humidity variations were greater. All parameters varied over time and among different heights and locations inside the greenhouses. Fan et al. [10] continuously monitored the microclimate inside solar greenhouses, analyzing temperature and humidity changes, and found a direct correlation between these factors and the greenhouses’ thermal efficiency. Wei et al. [11] characterized the CO₂ concentration inside a solar greenhouse. Three experiments were carried out to study changes in the CO₂ concentration and evaluate the impact of CO₂ enrichment on crop yield. The decrease in the CO₂ concentration in the greenhouse was more pronounced in the afternoon than in the morning, and ventilation had a major effect on the CO₂ concentration. In a study conducted by Ahamed et al. [12], a sensitivity analysis was carried out on a heating simulation model during winter. The study revealed that even small changes in indoor temperature and humidity had a significant impact on heating demand. It was found that extending the crop planting period required 13–20% more heat. Tadj et al. [13] also utilized heating techniques to investigate temperature and humidity distribution in greenhouses. Their findings demonstrated that heating pipes improved uniform climate conditions, while differences in temperature and humidity led to increased plant activity and crop transpiration.

Previous studies have primarily focused on variation in single environmental variables within greenhouses. However, our understanding of the impact of interactions between multiple environmental conditions on crop growth remains limited. This research sought to address this gap by exploring more effective environmental management strategies using precision agriculture technologies and considering their interactions. Reza et al. [14] described technologies that allow for real-time adjustments of critical variables such as temperature, humidity, and CO₂ concentrations. Nicolosi et al. [15] demonstrated the effectiveness of a flexible control system for managing greenhouse microclimates using a neural network approach. This system adjusted heating and cooling devices based on the predicted outside temperature, thereby enhancing the internal temperature to support crop production. Azaza et al. [16] proposed a fuzzy logic-based control system with wireless monitoring to track temperature, humidity, CO₂ levels, and light intensity. This system enabled control over internal environmental conditions, and with the application of sensor fusion across multiple points, it addressed environmental variability within the greenhouse economically [17]. Balendonck et al. [18] recommended a cost-effective wireless sensor network to analyze the horizontal distribution of temperature and humidity within greenhouses. Their findings revealed that long-term averaged spatial differences in temperature and humidity were 1.0–3.4 °C and 9–40%, respectively, although actual variations can be even larger. They also demonstrated that using at least 9 sensors per hectare (approximately 33 m apart) is necessary to reliably detect persistent cold or wet spots, especially in larger greenhouse environments. To ensure precise and reliable data, managing data across different nodes carefully is crucial, avoiding redundancy to meet the application’s requirements [19].

The integration and deployment of sensor technologies are essential for the precise control and management of environmental factors in a greenhouse. Depending on the variability of environmental factors, different sensors are required to optimize the monitoring and management of important environmental factors in CSGs. The objective of this study was to assess the spatial, vertical, and temporal variability in major environmental variables in two CSGs during the summer season, providing crucial insights to enhance environmental monitoring and optimal control practices.

2. Materials and Methods

2.1. Experimental Site

The study was conducted using two CSGs, each with dimensions of 60 m × 8 m (L × W). Both greenhouses were aligned along the east–west axis and situated at Shenyang Agricultural University in Liaoning, China (41°49′46″ N, 123°33′51″ E). One greenhouse was used to cultivate eggplant, while the other was left empty. Sunrise and sunset were at 4:21 and 19:04, respectively, with the weather conditions described as windy. The maximum average wind speed reaching 4.5 ms⁻¹ from the south–southeast at midday and by late afternoon, the wind direction shifted to south–southwest. In Northeast China, winter nighttime temperatures can fall below −25 °C, creating a significant challenge for sustaining a suitable thermal environment for plant growth in CSGs [20]. To enhance thermal performance in cold climates, the north wall of a CSG is often constructed using materials with both high thermal mass and insulation properties. Dense materials like brick can absorb and retain heat during the day, gradually releasing it at night to help stabilize the internal temperature. However, in practice, this effect is moderated by the inclusion of insulating materials, such as Styrofoam and air gaps, which reduce heat transfer, improving heat retention rather than dynamic heat storage [21].

The CSGs were constructed with a robust north-facing wall and a partially roofed north side, while the southern side was covered with protective material. The north wall, measuring 0.45 m in thickness, was composed of a layered brick structure that included Styrofoam insulation and an air gap to enhance thermal efficiency. This design was selected based on thermal performance simulations and structural requirements commonly recommended for cold-climate CSGs [22]. The combination of dense brick and lightweight polystyrene insulation leverages both thermal mass and high R-value properties, enabling the wall to absorb daytime heat and release it gradually at night [22]. Brick walls (density: 1800 kg/m³, specific heat capacity: 1050 J/kg·K) form the primary heat-storage layer, leveraging high thermal mass to absorb solar energy during the day [23]. Structurally, walls ≥ 0.4 m are also advantageous for supporting roof loads and resisting lateral forces from wind and snow, making this choice both practical and cost-effective.

The north roof, with a thickness of 0.2 m, was constructed from multiple layers of wood, Styrofoam, and various structural materials, providing additional insulation and structural integrity. Thermal modeling studies of CSGs in cold climates emphasize the importance of roof insulation, showing that increasing insulation thickness can reduce heat loss by up to 20%. The wood layers contribute to rigidity and load-bearing capacity, especially at the roof slopes of 32° (front) and 41° (rear), which are optimized for snow shedding and solar heat gain. The use of cost-efficient materials such as Styrofoam and wood also reduced construction expenses while ensuring durability.

The southern side of the greenhouse was covered with a polyvinyl chloride (PVC) film (0.00012 m thick), allowing for maximum sunlight penetration during the day. At night, a 0.50 m thick cotton blanket was deployed over the roof to enhance insulation. This dual-layer system—thin for light capture, thick for thermal retention—has been shown in CSG research to improve internal temperature regulation and reduce heating demands by 30–50% [24]. Together, the design choices for the wall and roof reflect a carefully considered balance of thermal performance, structural soundness, and cost-efficiency, consistent with best practices for energy-efficient greenhouse construction in high-latitude, cold-climate regions.

The greenhouse was equipped with a ventilation system, where the vent was opened at 06:30 and closed at 18:30 each day to regulate internal conditions. In one of the CSGs, eggplant (Solanum melongena L.) was planted on 5 April 2019. The soil beneath the plants was covered with a black plastic film to conserve moisture and control weed growth. The planting arrangement featured a plant-to-plant distance of 0.5 m and a row-to-row spacing of 0.25 m, with two rows per ridge. Each ridge was 6.70 m in length, and the spacing between adjacent ridges was 1.1 m. The plants had a height ranging from 0.30 to 0.40 m during the experimental period. In total, there were 40 ridges within the greenhouse, facilitating efficient use of space and resources for the cultivation of eggplants.

2.2. Wireless Sensor Network (WSN) and Sensors Used

To monitor the key environmental variables within the CSGs, this study used a comprehensive array of sensors as part of a wireless sensor network (WSN). The sensors provided continuous and precise measurements of various key parameters essential for understanding and optimizing the greenhouse environment, which are detailed in

Table 1. Specifications of the sensors used for monitoring environmental variables in CSGs.

Variable	Model	Range	Resolution	Accuracy
Temperature	AM2315	−20–80 °C	0.1 °C	±0.1 °C
Humidity	AM2315	0–100%	0.1%	2%
CO2	SH-300-ND	0–3000 ppm	1 ppm	2%
Light Intensity	BH1750FVI	0–65,535 lx	0.1 lx	±5%
Wind Speed	Davis 7911	0.5–89 ms−1	0.1 ms−1	±5%
Wind Direction	Davis 7911	0–360°	1°	7°

with the technical specifications. Temperature—humidity sensor (AM2315, Aosong Electronics Co., Ltd., Guangzhou, China) was used to monitor the air temperature and humidity within the greenhouses and outside. The sensor utilized a capacitive element and a wet component with a high-precision temperature unit integrated with a microprocessor for accurate and reliable temperature and humidity readings. The concentration of CO₂ was measured using a CO₂ sensor (SH-300-ND, SOHA TECH Co., Ltd., Seoul, Republic of Korea). To measure light intensity, both inside and outside the greenhouses, a light intensity sensor (BH1750FVI, Rohm Co., Ltd., Kyoto, Japan) was used with a principal photoresistor to obtain the amount of light. Additionally, an anemometer (Davis 7911, Davis Instruments Corp. Inc., Hayward, CA, USA) was used to detect the wind speed and direction outside the CSGs.

Within the WSN, the sensor nodes were interconnected with a central master node, facilitating the real-time transmission of the collected data to a central computing system at a frequency of 4 Hz [25]. A dedicated program was developed using a commercial software (LabVIEW, Version 2018, National Instruments, Austin, TX, USA) to streamline and optimize the data collection process. A Universal Asynchronous Receiver/Transmitter (UART) module (E15-USB-T2, Chengdu Ebyte Electronics Technology Co., Ltd., Sichuan, China) was used for wireless communication. This module housed the RF receiver and connected with the computer through serial communication [26]. Specifically, the communication range between these two wireless devices was capable of reaching up to 5 km [27]. This setup allowed for seamless and prompt transfer of the sensor data from the nodes to the central computer for further analysis and processing. Figure 1 illustrates the various components that make up the monitoring unit utilized in this study.

2.3. Procedure of Data Collection and Analysis of the Variability

Each greenhouse was segmented into three vertical planes for data collection, which were subdivided into three distinct layers. This stratification enabled a thorough spatial analysis of the environmental conditions at various heights. All measurements were conducted within three vertical planes inside the greenhouse. These planes were segmented into three horizontal layers: the bottom layer was positioned 0.50 m above the ground, the middle layer at 1.75 m, and the top layer at 3.0 m. The bottom layer height was selected to correspond closely with the plant canopy level (approximately 0.30–0.40 m), enabling effective monitoring of air temperature and humidity conditions within the plant microenvironment. Each layer within these planes was further subdivided into three specific sensor placement points, located 1.0 m, 4.0 m, and 7.0 m from the north wall of the greenhouse. The bottom and middle layers were each equipped with three temperature and humidity sensor nodes, while the top layer had a single sensor node located at the center point.

This configuration resulted in each vertical plane containing a total of seven temperature and humidity sensor nodes. Consequently, 21 sensors were deployed throughout the entire greenhouse, providing comprehensive coverage and precise monitoring of the environmental conditions within the structure. A CO₂ sensor and a light intensity sensor were installed at the center of each layer in the middle vertical plane of the greenhouse. This configuration resulted in a total of three CO₂ sensors and three light intensity sensors being used within each greenhouse. Additionally, five external sensors were installed 5 m above and outside the greenhouse to measure air temperature, humidity, wind speed and direction, solar radiation, and light intensity. Figure 2A illustrates the placement of sensor nodes within the greenhouse for internal data collection across different cross-sections, as well as the positioning of external sensors. Figure 2B provides a detailed view of the distances between the sensors and their respective heights within a cross-section of the greenhouse. All sensors were installed on 21 May 2019, with data collection commencing on the morning of the installation day.

The WSN continuously recorded data from both inside and outside the greenhouse throughout the experiment. Each sensor node consisted of an Arduino UNO, connected sensors, power supplies, and UART wireless transceiver modules. The network was structured with transmitter nodes, known as slaves, which sent data to the receiver nodes, referred to as masters. After assembling the system, the sensor nodes underwent calibration, as even sensors from the same manufacturer can produce varying readings. Calibration involved taking measurements over approximately 15 min under controlled temperature and humidity conditions, with three repetitions to ensure accuracy. During the experiment, the spatial and vertical variability of the environmental variables was analyzed to assess their distribution patterns. This analysis was crucial in determining the optimal placement of the sensors within the greenhouse. Additionally, the temperature and humidity data were segmented by time for further analysis [28].

To visualize the spatial and temporal distribution of temperature and humidity within the greenhouses, the environmental data collected from the wireless sensor network were segmented by specific time intervals throughout the day. For each selected time point, data from all relevant sensor nodes within each greenhouse were compiled. Spatial interpolation was then performed using the Inverse Distance Weighting method via ArcGIS software (version 10.8.1; Esri, Redlands, CA, USA). This geostatistical method estimates values at unsampled locations based on the values of nearby measured points, with closer points weighted more heavily. The interpolation process was applied to each vertical plane (front, middle, rear) and for each horizontal layer (bottom, middle, top) separately. The resulting interpolated surfaces were used to construct the temperature and humidity maps. These maps effectively illustrate the spatial and vertical variability of environmental conditions in both the greenhouse with crops and the one without, providing a visual representation of microclimate dynamics at different times of the day.

2.4. Determination of Optimal Placement of the Sensors

To effectively manage the internal environment of a greenhouse, it is essential to strategically position sensors to accurately reflect the overall conditions. In this study, an error-based approach was used to determine the optimal sensor placement locations. This method identified the combination of sensor positions that most accurately represented the environmental conditions throughout the entire greenhouse. This methodology was implemented under the assumption that an accurate representation of the overall environment can be achieved by averaging the data collected from all sensors. The process of determining the optimal sensor locations involved an error-based method [8]. This approach was designed to minimize the discrepancy between a reference trend, which was derived by averaging data from all sensors, and various combination trends, which were calculated by averaging data from different sensor location combinations. The aim was to determine the locations that would collectively provide the most accurate representation of the entire greenhouse environment through this error minimization approach. According to the statistical indicator, the variations between the reference trend and combination trends were examined according to the number of sensors installed. The root mean square error (RMSE) was used as the evaluation index.

To support the conclusions regarding environmental differences and optimal sensor placement, descriptive and inferential statistical analyses were conducted. Descriptive statistics included means and standard deviations to summarize overall trends. One-way analysis of variance (ANOVA) was used to identify significant differences in environmental parameters across different layers and zones within the greenhouses. Where significant effects were detected, Tukey’s Honest Significant Difference (HSD) test was applied for pairwise comparisons to determine the specific locations with statistically distinct conditions. All statistical analyses were performed using Statistix 10 (Analytical Software, Tallahassee, FL, USA). These statistical methods provided robust validation of the observed spatial patterns and supported the determination of optimal sensor locations for accurate environmental monitoring.

3. Results

3.1. Variation in Light Intensity

Light intensity significantly affects plants’ growth and development. Light intensity directly influences photosynthesis, a crucial process for plant growth. Higher light intensity generally increases the rate of photosynthesis, leading to better growth and higher yields. However, excessively high light intensity can cause photoinhibition, damaging the plants’ tissues. Figure 3 and Figure 4 show the average external light intensity and the light intensity of different layers inside both greenhouses.

The highest light intensity was observed around midday (12:00) both outside and inside the greenhouse. The external light intensity reached approximately 57,000 lux, whereas inside the greenhouse, it was around 54,000 lux across all layers (bottom, middle, and top). Light intensity began to rise sharply after 08:00, peaking at noon, and then declining steadily after 14:00. It dropped to nearly zero from 20:00 until around 06:00 the next day. Inside the greenhouse, the light intensity was relatively uniform across the bottom, middle, and top layers, although the bottom layer showed a slightly lower intensity than the middle and top layers. This indicates effective distribution of light, albeit with room for improvement at the lower levels.

The graphs clearly show that light intensity inside the greenhouse was slightly lower than that outside, with a peak intensity of 54,000 lux inside compared with 57,000 lux outside. This slight reduction in light intensity can impact photosynthesis, which is directly related to plant growth. Photosynthesis rates increase with light intensity up to a certain point, meaning that the higher the light intensity, the more energy plants can convert for growth and development. However, the relatively uniform light intensity across the different layers inside the greenhouse suggests that plants at various heights receive adequate light, promoting balanced growth [29]. The slight reduction in light intensity at the bottom layer compared with the middle and top layers indicates that plants at lower levels may receive less light, potentially affecting their growth rate and health. Therefore, ensuring that light reaches all layers effectively is crucial for maximizing overall plant growth and yield.

To optimize light distribution and ensure that all plants receive sufficient illumination, several strategies must be considered. The use of reflective materials on the walls and floor of the greenhouse can enhance the distribution of light to the lower layers by reflecting light, thereby minimizing light loss [30]. This approach may help to reduce the differences in the light intensity. Installing light-diffusing panels can scatter the incoming light, modifying shadows and promoting the uniform distribution of light across the entire plant canopy. These panels facilitate deeper penetration of light into the plant layers, benefiting those situated at the bottom. Properly spacing plants and arranging them to minimize shading can enhance the penetration of light to the lower leaves. Ensuring that taller plants do not overshadow shorter ones can help maintain consistent light levels across all layers. Implementing adjustable lighting systems that can be repositioned closer to the plants during low-light periods ensures all plants receive adequate light. Such systems are particularly effective in evening out the distribution of light across varying plant heights. Additionally, using supplemental lighting, such as LED lights, can help maintain consistent light levels during periods of low natural light [31,32]. LEDs can be strategically positioned to target areas with lower light intensity, such as the bottom layer, ensuring that all plants receive sufficient light for optimal growth.

3.2. The Variability of Temperature and Humidity

Figure 5 and Figure 6 illustrate the overall variations in temperature and humidity recorded both inside and outside two distinct greenhouses (with and without crops) during the experimental period. The minimum and maximum temperatures outside the greenhouses were observed at 32 °C at 12:00 and 12 °C at 02:00, respectively. The relative humidity outside the greenhouses ranged from a minimum of 16% at 12:00 to a maximum of 100% between 16:00 and 22:30. The temperature showed a consistent increase from 08:00, reaching its peak around 13:00, after which it began to decline. In contrast, the relative humidity decreased from 14:00 to 12:00, before increasing steadily until 08:00. Inside the greenhouse without crops, the average minimum temperature was recorded at 17 °C at 08:00, while the maximum temperature reached 40.2 °C at 12:00. The relative humidity inside this greenhouse varied, with a minimum of 25.25% at 12:00 and a maximum of 45% at 16:00. Similarly, within the greenhouse containing crops, the internal temperature ranged from an average minimum of 17 °C at 08:00 to a maximum of 40.2 °C at 12:00. The relative humidity within this greenhouse followed a similar pattern, with a minimum of 25.25% at 12:00 and a maximum of 45% at 16:00.

Figure 5 presents the temperature variations across different layers in both greenhouses (with and without crops). In the greenhouse without crops, the temperature difference between the layers was approximately 5.2 °C, while in the greenhouse with crops, the temperature difference was around 4.8 °C. The bottom layer of the greenhouse containing crops was warmer compared to that in the greenhouse without crops, while in both cases, the top layer maintained a higher temperature than the bottom layer. The overall average temperature inside both greenhouses was approximately 17 °C, which was higher than the average outside temperature. Figure 6 illustrates the differences in humidity between layers in both greenhouses. In the greenhouse without crops, the humidity difference between layers remained lower during the daytime, with smaller variations between layers compared to the greenhouse with crops. However, during the night, the humidity differences became more pronounced, with the lower layer exhibiting higher humidity than the top layer. In contrast, the greenhouse with crops exhibited more significant humidity differences between layers throughout the day and night. During the daytime, the bottom layer had lower humidity than the top layer, likely due to the increased crop respiration rate during this period. At night, however, the humidity of the bottom layer was higher than that of the other layers. In both greenhouses, the inside humidity was approximately 25% lower than the outside humidity. Despite the occurrence of rainfall around noon during the experimental period, it had a minimal effect on the internal temperature conditions of the greenhouses. Overall, the data indicate that temperature and humidity inside the greenhouses varied significantly across different times of day and at different heights, providing important insights into the microclimatic conditions that influence greenhouse management and crop performance.

The experimental data from both solar greenhouses (with and without crops) from three different sides (front, middle, and rear) were analyzed. Maps of the temperature and humidity for both greenhouses (with and without crops) were made for various heights using this procedure. Maps of the temperature and humidity were shown at specific hours during the day to display the spatial and temporal variations in the temperature and humidity for each greenhouse. In Figure 7 and Figure 8, the average internal temperatures of the CSGs with and without crops are shown for the front, middle, and rear at different heights. Figure 9 and Figure 10 show the average internal humidity of the CSGs with and without crops for the front, middle, and rear at different heights. The temperatures were higher during the daytime in both greenhouses. This indicated that the temperature increased during the daytime, and the temperature of the top layer was higher than that of the bottom layer inside the greenhouse without crops. In addition, the motion of air within the greenhouse was restricted by the presence of crops; consequently, the bottom layer with crops was warmer than the bottom layer without crops. Moreover, the air temperature in both greenhouses decreased when the sun set. The brick wall radiated the stored energy during the night and kept the greenhouse warm, although the bottom layer was cooler than the middle and top layers. Rain had a small effect on the overall distribution of temperature in both greenhouses. The humidity was lower during daylight hours than at night in both greenhouses. Humidity was as high as 85% at night in both greenhouses. The distribution of humidity was more stable in the greenhouse with crops than in the greenhouse without crops. During the summer, the temperature and humidity increased and decreased significantly. Monitoring the temperature and humidity using sensors is critically important for determining how the environment inside greenhouses could be best controlled to enhance crop growth and productivity.

The observed temperature and humidity patterns inside both greenhouses are closely linked to the ventilation schedule, which operated daily from 06:30 to 18:30. During this period, external air was introduced into the greenhouse environment, helping to dissipate accumulated heat and reduce internal humidity. This effect is particularly evident in the decline of temperature after noon and the minimum humidity levels recorded around 12:00 in both greenhouses. The ventilation helped moderate extreme microclimatic conditions during the hottest part of the day, contributing to thermal regulation. Conversely, after 18:30, when the vents were closed, internal temperature and humidity levels began to stabilize or rise due to reduced air exchange and continued transpiration and respiration. These observations suggest that ventilation not only aids in passive cooling but also directly influences humidity stratification and moisture retention within the greenhouse. Therefore, integrating ventilation control with real-time environmental monitoring is essential for maintaining optimal growing conditions and improving crop productivity in CSG.

3.3. Variation in CO₂

The average maximum and minimum CO₂ concentrations outside the greenhouses were 546 ppm at 10:00 (earlier in the measurement period) and 449 ppm at 08:00 (near the end of the 24 h cycle), respectively. Inside the greenhouse without crops, the average maximum and minimum CO₂ concentrations were 1000 ppm at 02:00 and 571 ppm at 08:00 at the top layer. Within the greenhouse containing crops, the average maximum and minimum CO₂ concentrations were 1020 ppm at 02:00 and 556 ppm at 08:00 at the middle and top layers, respectively.

Plants require CO₂ for photosynthesis, and the optimal CO₂ level varies depending on the plant species and light conditions in the greenhouse [33]. As light intensity decreases in the greenhouse, both photosynthesis and consumption of CO₂ decrease [34]. Figure 11 illustrates the CO₂ concentrations outside the greenhouses and inside the greenhouse containing crops at different layers. The highest CO₂ concentration was at night due to plant respiration. Conversely, the CO₂ concentration within the greenhouse without crops showed little variation between day and night, as shown in Figure 12.

CO₂ enrichment is crucial in modern greenhouse management, significantly affecting crop yields, photosynthesis, and overall plant growth. By artificially increasing the CO₂ concentrations in greenhouses, farmers can enhance plants’ growth rates and yields. Optimal CO₂ levels for most greenhouse crops typically range between 800 and 1000 ppm, with a peak of 1020 ppm observed in greenhouses with crops, which fell within this beneficial range [35]. Elevated CO₂ levels improve photosynthetic efficiency, resulting in faster growth rates and higher yields, particularly for crops such as eggplant. Moreover, CO₂ enrichment can enhance water use efficiency in plants, enabling them to maintain growth even with reduced water availability.

However, the CO₂ dynamics within greenhouses demonstrate marked daytime variations. During daylight hours, plants actively consume CO₂ for photosynthesis, leading to a significant drop in CO₂ levels, especially during periods of high light intensity. When photosynthesis ceases, plants respire and release CO₂ at night, causing higher concentrations. This pattern is evident in the results, with the highest CO₂ concentrations inside greenhouses recorded at night (up to 1020 ppm at 02:00) and the lowest in the morning (down to 556 ppm at 08:00). This fluctuation highlights the necessity for effective CO₂ management strategies that account for these natural variations to optimize plants’ health and productivity.

These data emphasize the importance of continuous monitoring and effective management of the CO₂ levels in CSGs. Significant variations in CO₂ concentrations throughout the day and night necessitate strategies that dynamically adjust to these changes, optimizing plants’ growth and yield. Greenhouse operators can leverage this information to better understand and implement CO₂ management practices, ultimately enhancing crop yield, promoting healthy plant growth, and maintaining optimal greenhouse environments. Through meticulous management of CO₂, light, and other environmental factors, farmers can achieve sustainable and productive CSG operations.

3.4. Locations of the Temperature and Humidity Sensor for Optimum Monitoring and Control

The analysis included both the ANOVA test and Tukey’s HSD test to assess whether the measured data exhibited statistically significant differences across various locations of the sensors. The distribution of temperature and humidity at each location is visually represented in the box plots depicted in Figure 13 and Figure 14. Different letters positioned above the box plots signify significant differences at p < 0.05 according to Tukey’s HSD test. The results of ANOVA indicated a statistically significant difference in temperature across different locations within the greenhouse. Subsequent analysis using Tukey’s HSD test allowed the categorization of locations into three distinct groups for the conditions both with crops and without crops.

The findings suggest that without a crop, the front and middle layers exhibited differences compared with other conditions, implying that optimal monitoring necessitates the use of all 21 sensors. However, in real greenhouse experiments involving crops, the statistical analysis revealed no significant differences among the front, middle, and rear layers. Consequently, for optimal placement of sensors in a greenhouse with crops, a single layer, with only seven sensors, would be sufficient for comprehensive monitoring rather than the full set of 21 sensors. This insight emphasizes the efficiency and resource optimization in the deployment of sensors for environmental monitoring of solar greenhouses during actual crop cultivation scenarios in summertime.

While the ANOVA and Tukey HSD tests identified a single layer comprising seven sensors to be sufficient for monitoring the solar greenhouse, determining the specific group of sensors for optimization remained challenging. To address this issue, the error-based method was subsequently used to determine the most accurate positions for the sensor to optimize and effectively monitor the solar greenhouse. A comparison was conducted between the temperature and humidity data obtained from various groups of sensors and the average data collected from all sensors. The RMSE was used as a measure to evaluate the data recorded by each sensor relative to the reference trend. These comparative results are presented in

Table 2. The root mean square error (RMSE) as a metric for assessing the accuracy of the data measured by each group of sensors in comparison with the reference trend.

	Temperature (°C)						Humidity (%)
	With Crops			Without Crops			With Crops			Without Crops
	F *	M *	R *	F *	M *	R *	F *	M *	R *	F *	M *	R *
Average	25.9	24.5	24.3	25.9	24.4	24.3	51.5	51.8	47.3	55.9	56.0	49.9
RMSE	0.9	0.3	0.6	1.0	0.4	0.6	1.3	1.6	2.9	1.9	2.1	4.0

* F = Front layer, M = Middle layer, R = Rear layer.

.

In Table 2, it is observed that the RMSE value of temperature for the middle layer is lower than that of any other layer. This suggests that the data measured in the middle layer closely align with the reference trend for temperature. Conversely, for humidity, the front and middle layers presented smaller RMSE values. Traditionally, sensors are placed in the middle of a greenhouse for environmental control. However, environmental data collected from the middle may not consistently represent the entire greenhouse environment accurately. Therefore, the selection of the sensors’ locations should be based on the greenhouse’s design and control strategies rather than a default central placement.

Summing up the analysis, it can be concluded that using only the group of sensors in the middle layer would be sufficient for effective monitoring. To improve accuracy and cost-effectiveness, farmers might integrate sensor clusters in the front and middle layers. While Figure 7, Figure 8, Figure 9 and Figure 10 indicate that the greenhouse environment is spatially varied, these locations were determined through an error-based optimization process that offers the best balance between accuracy and deployment costs, effectively reflecting the overall greenhouse conditions.

4. Discussion

The present study provides valuable insights into the microclimatic conditions within the CSGs with and without crops. The results reveal significant variations in temperature, humidity, CO₂ concentration, and light intensity based on temporal and spatial factors. These findings align with previous studies emphasizing the dynamic nature of greenhouse environments and the necessity of precise monitoring and control strategies to optimize plant growth and productivity [9,36].

One of the key observations was the variation in temperature within the greenhouses. The temperature inside both greenhouses was consistently higher than the outside temperature, with notable differences in the top and bottom layers. This supports the findings of [1,2], who reported that stratification within a greenhouse can impact heat distribution and crop growth. The presence of crops in the greenhouse influenced temperature and humidity distribution, as plants modified the microclimate through transpiration and shading effects [13,37]. The data indicate that the top layer maintained a higher temperature than the bottom layer, which is expected due to heat accumulation near the roof. The increased bottom layer temperature in the greenhouse with crops can be attributed to restricted air movement and heat retention by the soil, as highlighted by Li et al. [38]. The humidity variations observed in this study are consistent with the findings of prior research, which suggest that greenhouses with crops exhibit greater humidity fluctuations due to plant transpiration [39]. The fluctuations highlight the necessity of effective ventilation and humidity control to prevent excessive moisture accumulation that can lead to fungal diseases [40].

The temperature and humidity distribution maps in this study reveal clear spatial and temporal patterns consistent with established microclimate behaviors in CSGs. As reported by Tong et al. [2] and Guohong et al. [3], diurnal fluctuations were observed, with daytime heating and nighttime cooling dominating internal conditions. A key finding of this study is the demonstration of the moderating effect of crops on microclimate variability. Consistent with Liang et al. [4] and Singh et al. [7], vegetation enhanced thermal stability, particularly in the lower layers, by limiting air circulation and increasing heat retention through transpiration. The crop-filled greenhouse showed higher bottom-layer temperatures and more consistent humidity compared to the empty structure. The thermal storage role of the north brick wall, which released heat during the night, aligns with Wang et al. [1], highlighting the importance of passive heating in maintaining nighttime temperatures.

Ventilation significantly influenced microclimate dynamics. Daytime ventilation (06:30–18:30) reduced internal temperatures and humidity, with the lowest humidity occurring around noon—a pattern also observed by Qiu et al. [5] and Ryu et al. [9]. After vent closure, temperature and humidity gradually increased, supporting the case for adaptive, sensor-driven ventilation systems as proposed by Luo et al. [6] and Nicolosi et al. [15]. The observed spatial heterogeneity aligns with findings by Ferentinos et al. [17] and Balendonck et al. [18], reinforcing the need for precise sensor placement across vertical and horizontal zones to ensure accurate monitoring. This study enhances current understanding of CSG microclimates by combining spatial–temporal analysis with statistical validation. It highlights the critical roles of crop presence, structural thermal mass, and responsive ventilation in optimizing growing conditions in cold climates.

The vertical CO₂ concentration profiles highlight the strong influence of the ventilation schedule on greenhouse gas dynamics. During active ventilation (06:30–18:30), CO₂ levels declined—particularly between 08:00 and 14:00—due to photosynthetic uptake and dilution from air exchange. After the vents closed at 18:30, CO₂ levels increased overnight from plant and soil respiration. The top layer consistently showed higher daytime CO₂ concentrations, likely due to thermal convection and vent positioning affecting air mixing. These findings indicate the need to align ventilation timing with plant physiological activity to optimize internal CO₂ levels for growth.

The observed variations in CO₂ concentration within the greenhouse align closely with previously reported patterns in the literature, reflecting the complex interplay of plant physiological processes and environmental controls. Higher nighttime CO₂ concentrations observed in this study, reaching up to approximately 1020 ppm, are consistent with findings by Wei et al. [11], where nocturnal CO₂ accumulation is attributed to plant and microbial combined with limited ventilation. Interestingly, the measured concentrations consistently fell within the optimal CO₂ range (800–1000 ppm) recommended for enhanced photosynthetic efficiency and crop yield, as described in previous greenhouse management studies [35,41]. This alignment with optimal CO₂ ranges reinforces the importance of controlled ventilation timing in maintaining beneficial nighttime CO₂ accumulation. However, the observed significant daytime reductions in CO₂ concentration suggest potential areas for improvement, particularly through targeted CO₂ enrichment strategies during daylight hours, as proposed by prior research [38]. Incorporating automated CO₂ management systems could optimize CO₂ levels throughout the diurnal cycle, potentially leading to increased growth rates and water-use efficiency, an approach successfully demonstrated in other greenhouse studies [35,42]. Thus, our findings not only confirm the existing literature but also highlight opportunities for refining greenhouse management practices to leverage natural and supplemental CO₂ dynamics effectively.

The analysis of sensor placement using statistical techniques, including ANOVA and Tukey’s HSD test, revealed that while all 21 sensors were necessary for an empty greenhouse, only seven sensors were required when crops were present. This finding suggests that optimal sensor placement should be tailored to the greenhouse environment rather than relying on uniform distribution [14,25,43]. These results emphasize the importance of site-specific sensor placement to optimize environmental monitoring and control strategies in greenhouses. While the initial analysis confirmed that deploying all 21 sensors may be necessary for comprehensive monitoring in an empty greenhouse, the statistical evaluation (ANOVA and Tukey’s HSD test) indicated that, in practical cultivation scenarios with crops, environmental conditions become more uniform across vertical zones. As a result, only a single horizontal layer, comprising 7 strategically placed sensors, is sufficient to accurately monitor environmental conditions. This reduction in sensor count represents a significant decrease in hardware, installation, and maintenance costs, improving the feasibility of large-scale implementation in commercial greenhouses [17,19].

To further refine the sensor deployment strategy through an error-based optimization approach, it was identified that the middle layer yielded the lowest RMSE for temperature, while both front and middle layers were optimal for humidity. A hybrid configuration using these zones was found to offer the best balance between data accuracy and system cost. This result aligns with findings by Balendonck et al. [18], who emphasized the importance of spatial variability in determining sensor locations for resource-efficient greenhouse monitoring. In addition, studies like Wang et al. [1] and Iqbal et al. [25] supported the integration of optimized sensor networks with intelligent control systems, noting their potential in improving climate control while reducing energy input and operational costs. The recommended sensor network configuration enables precision monitoring with reduced economic burden, supporting sustainable and scalable deployment.

Despite the valuable insights provided by this study, several limitations must be acknowledged. First, the study was conducted over a limited experimental period, which may not capture seasonal variations in greenhouse microclimate conditions. Future research should extend the study period to analyze long-term trends across different seasons. Additionally, the study focused on a single crop type; therefore, further investigations should explore variations in microclimatic conditions for different crops with varying canopy structures and transpiration rates. Another limitation was the lack of real-time adaptive control mechanisms, such as automated ventilation and CO₂ supplementation, which could enhance greenhouse climate management. Future research should incorporate intelligent greenhouse control systems integrating artificial intelligence and IoT-based sensor networks to optimize environmental conditions dynamically. Finally, while this study analyzed temperature, humidity, CO₂ concentration, and light intensity, other crucial factors such as soil moisture and nutrient levels should also be examined to provide a comprehensive understanding of greenhouse crop production. Addressing these limitations in future studies will further refine greenhouse management strategies, leading to improved crop productivity and resource efficiency.

5. Conclusions

This experiment on winter season CSGs provided key insights into environmental dynamics and optimal cultivation conditions. Using a wireless sensor network, the study monitored temperature, humidity, CO₂ levels, and light intensity, enabling detailed spatial and temporal analysis. The greenhouse with crops exhibited slightly lower vertical temperature differences (4.8 °C) compared to the greenhouse without crops (5.2 °C), and the bottom layer was warmer in the presence of crops. Humidity distribution was more stable in the greenhouse with crops, suggesting that crop transpiration and physical structure help regulate internal microclimatic conditions. The presence of crops modified the vertical temperature and humidity distribution by restricting air movement, which led to warmer bottom layers in the greenhouse with crops compared to the one without. Daytime temperatures were consistently higher, and nighttime cooling was mitigated by the thermal mass of the brick wall in both greenhouses. In the greenhouse with crops, CO₂ concentrations peaked at night (1020 ppm at 02:00) due to plant respiration and dropped in the morning (556 ppm at 08:00) when photosynthesis began. In contrast, the greenhouse without crops exhibited minimal diurnal variation in CO₂ levels, indicating that crop presence plays a key role in modulating internal CO₂ dynamics. Statistical tests confirmed that in greenhouses with crops, there were no significant spatial differences (front, middle, rear), allowing a reduction from 21 to just 7 sensors (one layer) without compromising monitoring accuracy. This finding highlights a cost-effective and resource-efficient strategy for environmental sensing during actual crop cultivation.

This study effectively monitored air temperature, humidity, and CO₂ concentration in summer season CSGs, but lacked soil temperature data—an important limitation given the influence of ground coverings like black plastic mulch on the microclimate. Future studies should include soil sensors to capture a more comprehensive plant environment. The findings confirm that optimal sensor placement, particularly using error-based RMSE analysis, enhances climate monitoring accuracy. The proposed framework supports improved CSG management through strategic sensor deployment, crop arrangement, CO₂ enrichment, and thermal storage integration. Incorporating advanced sensor networks and analytics will further optimize productivity and environmental control in hot-climate greenhouses.