Research Status and Prospects on the Construction Methods of Temperature and Humidity Environmental Models in Arbor Tree Cuttage

Abstract

The environmental temperature and humidity are crucial factors for the normal growth and development of arbor tree cuttings by altering their hormone levels and influencing their physiological activities. Developing a temperature and humidity environmental model for arbor tree cuttings serves as a key technique to improve the adjustment performance of environmental parameters in facility agriculture systems and enhance the rooting rate of cuttings. This paper provides a comprehensive summary of current research on the inherent characteristics of cuttings and the factors influencing environmental temperature and humidity. It explores the mechanisms of interaction between the inherent characteristics of cuttings and the factors influencing environmental temperature and humidity. This paper investigates the interactive relationships among the factors affecting environmental temperature and humidity. It analyzes methods to improve the efficiency of constructing temperature and humidity environmental models for arbor tree cuttings. To enhance the transferability of the environmental model, the necessary physiological activities under the influence of plant hormones are generalized as common physiological traits in the growth and development of cuttings. In addition, this paper explores the factors influencing the air and substrate temperature and the humidity in facility agriculture systems as well as two types of facilities for controlling environmental temperature and humidity. Furthermore, it reviews the research progress in environmental models from both mechanistic and data-driven perspectives. This paper provides a comparative analysis of the characteristics associated with these two model categories. Building upon this, the paper summarizes and discusses methods employed in constructing temperature and humidity environmental models for arbor tree cuttings. In addition, it also anticipates the application of deep learning techniques in the construction of temperature and humidity environmental models for arbor cuttings, including utilizing machine vision technology to monitor their growth status. Finally, it proposes suggestions for building physiological models of fruit tree-like arbor cuttings at different growth stages. To enhance the transferability of environmental models, the integration of physiological models of cuttings, environmental models, and control system performance are suggested to create an environmental identification model. This paper aims to achieve control of the common physiological activities of cuttings.

1. Introduction

Cuttage is a technique in plant in vitro organ regeneration. After cuttings adapt to the temperature and humidity environment, their tissue cells undergo dedifferentiation into totipotent cells. These totipotent cells then re-differentiate to form the root primordium, which develops into adventitious roots, ultimately growing into new plants [1,2]. Cutting seedling cultivation boasts the advantages of a short cycle and the preservation of excellent maternal traits. It can provide outstanding germplasm for molecular-assisted breeding. This technology plays a crucial role in the exploration of genes determining essential genetic traits and the enhancement of breeding efficiency [3,4]. Before the cuttage seedlings of arbor trees are industrialized or cuttage seedling trials for new arbor varieties are conducted, the environmental temperature and humidity conditions suitable for the cuttings can only be obtained after prolonged monitoring and recording. Researchers subsequently based the research on the growth status of the cuttings and combined the production experience to summarize the environmental parameters needed for this cutting type. This process serves as the basis for setting parameters in the environmental control system. It is time-consuming and involves numerous influencing factors, and it is leading to a certain degree of randomness and subjectivity in the construction of the environmental model. To improve the efficiency of constructing environmental models for arbor tree cuttings, it is necessary to summarize the physiological commonalities exhibited by the inherent characteristics during the rooting process of cuttings as well as the mechanisms of interaction among the factors influencing environmental temperature and humidity. This helps to enhance the transferability of the environmental model, enabling it to adapt to the process of new varieties and consequently achieve efficient production and research.

The inherent characteristics of cuttings are composed of both their external traits and internal physiological activities since cuttings come from the mother tree. The quality and type of the mother tree directly impact the physicochemical properties of the cuttings [5]. The location of the cutting slips significantly influences the survival rate and the nursey emergence rate of the cuttings [6]. The length of cuttings determines the growth effectiveness of their adventitious roots [7]. Additionally, the number of leaves on the cutting branches is one of the factors influencing root growth [8]. The tissue cells of cuttings undergo necessary physiological activities under the influence of hormones to complete their own growth and development, including photosynthesis, transpiration, and respiration. These cellular activities can be considered as physiological commonalities during the rooting process of cuttings. Therefore, the proportion of plant hormones in cuttings is a key factor influencing the rooting rate, and the content of plant hormones is closely related to the environment in which cuttings are planted.

There are numerous factors influencing the rooting rate in cuttage circumstances, but temperature and humidity directly impact the synthesis, distribution, and transportation of hormones in cutting cells, thereby indirectly altering the growth status of cuttings [9,10]. Therefore, facility agriculture systems with precise environmental temperature and humidity control capabilities are more conducive to experimental research on environmental parameters for arbor tree cuttings. Facility agriculture systems, including plant factories, greenhouses, artificial climate chambers, etc., are comprehensive projects that integrate control theory, meteorology, botany, and other disciplines [11]. Within these agriculture systems, artificial climates can be created to address the intricate environmental requirements for plant growth. Both domestically and internationally, scholars have conducted extensive research on facility agriculture systems, yielding enriched achievements [12]. As early as the 1960s, European scholars initiated research on the production and environmental monitoring techniques of facility agriculture systems [13]. The control methods gradually evolved into multi-factor collaborative control [14], with hardware transitioning from single-chip microcontroller control [15] to IoT-based control [16]. The current advancements in artificial intelligence technology [17] have propelled the environmental monitoring technology of facility agriculture systems into a new phase of rapid advancement [18].

By employing artificial intelligence technology to construct temperature and humidity environmental models for the rooting process of arbor tree cuttings, researchers can monitor factors influencing environmental temperature and humidity, which allows them to gather representative data for training the model. The facility agriculture system is categorized into two areas based on covering materials involved in the internal environment and the external environment. Furthermore, based on the presence or absence of a substrate, the internal environment is subdivided into the substrate environment and the air environment. Therefore, this paper begins with providing an overview of the inherent characteristics of arbor tree cuttings. It then reviews the factors influencing the internal environment’s temperature and humidity in facility agriculture systems, respectively. Subsequently, it outlines temperature and humidity control facilities in facility agriculture systems and methods for building environmental models. Finally, the paper concludes with a summary and outlook on the methods of building temperature and humidity models for cuttings, providing additional recommendations for building models of fruit trees. The method for constructing the cuttings’ temperature and humidity environmental model is illustrated in Figure 1.

2. The Characteristics of Arbor Plant Cuttings

2.1. External Characteristics of Cuttings

The external characteristics of cuttings are mainly manifested in the external form and structure. Based on the degree of lignification, cuttings can be divided into shoots [19] and hard wood [20,21]. The number of buds on the cuttings of the two types is different, and the buds of shoots are generally more than that of hard wood. Before cuttings, it is required to ensure that the top opening is flush and the bottom opening is cut at 45°. During cuttings, appropriate depth is needed for operation, and the cutting substrate should be gently compressed after cuttings so as to ensure that as many tissue cells contact the substrate as possible and increase the number of callus cells that can be generated [22]. The rooting process of apple rootstock shoots is an example for illustration, as shown in Figure 2. Eight typical tree cuttings are selected to describe the cutting parameters, as shown in Table 1.

As indicated in Figure 2, it can be seen that the tissue cells of the apple rootstock cutting slips underwent dedifferentiation to produce a callus, at which time the cells in the callus were totipotent, and after redifferentiation into root primordia, adventitious roots were eventually produced, after which the number and length of the adventitious roots increased continuously [23].

Table 1. Parameters of 8 kinds of arbor plants cuttings.

Cuttings Type	Notch Angle (°)	Length (cm)	The Number of Buds (Piece)	The Number of Leaves (Piece)	Depth of Insertion into Substrate (cm)	Rooting Rate (%)	Related Research
Salix Mongolica	45	100	3~5	2	66.7	97.92	Zha et al. [6]
Fir	45	15	1	Retain the Needle Leaves Above 2 cm at the Lower End	2	85.83	Chen et al. [7]
Toona Ciliata	45	10		2	2~3	94.44	Chen et al. [5]
Birchleaved Pear	45	10~15	1	3~5	3.33~5	70	Wang et al. [24]
Cherry Rootstock (Gisela)	45	10~12		1~3	3.33~4	85	Ren et al. [25]
Peach Rootstock (GF677)	45	10~15	2~3	3~4	3~4	83	Tewfik [26]
Apple Rootstock (Laoshannaise)	45	15~20	3~4	3~4	1.5~2	90	Xiao [27]
Mountain Apricot	45	8~10		2~3		66.7	Dai et al. [28]

Table 1. Parameters of 8 kinds of arbor plants cuttings.

Cuttings Type	Notch Angle (°)	Length (cm)	The Number of Buds (Piece)	The Number of Leaves (Piece)	Depth of Insertion into Substrate (cm)	Rooting Rate (%)	Related Research
Salix Mongolica	45	100	3~5	2	66.7	97.92	Zha et al. [6]
Fir	45	15	1	Retain the Needle Leaves Above 2 cm at the Lower End	2	85.83	Chen et al. [7]
Toona Ciliata	45	10		2	2~3	94.44	Chen et al. [5]
Birchleaved Pear	45	10~15	1	3~5	3.33~5	70	Wang et al. [24]
Cherry Rootstock (Gisela)	45	10~12		1~3	3.33~4	85	Ren et al. [25]
Peach Rootstock (GF677)	45	10~15	2~3	3~4	3~4	83	Tewfik [26]
Apple Rootstock (Laoshannaise)	45	15~20	3~4	3~4	1.5~2	90	Xiao [27]
Mountain Apricot	45	8~10		2~3		66.7	Dai et al. [28]

The survival of cuttings is reflected by the characteristics of the roots. The more developed the roots are, the better the cutting effect is. Yoon et al. studied the growth and development of salix mongolica cuttings and took the maximum root length of cuttings as the standard for evaluating the rooting effect [29]. Dawa et al. studied the root generation process of plant cuttings and reflected the cutting effect with the average root length of the cuttings [30]. Kalanzi et al. studied the effect of the segment cutting position on the rooting of cuttings and showed the significance of the treatment by analyzing the amount of rooting [31]. The root weight and the number of root segments are positively correlated with the growth rate of cuttings [32]. When studying the cutting effect of each index, the mean values of the affiliation function values of the different determinants under each treatment were taken as the evaluation standard. The larger the value of the affiliation function is, the more effective the cuttings are. The index measurement formula is as follows:

$$R(X_j)=(X_j-X_{\min })/(X_{\max }-X_{\min })$$

(1)

X_j—A processed indicator value;

X_max—Index maximum;

X_min—Index minimum.

It can be calculated from Table 1 that the average rooting rates of non-fruit trees and fruit trees are 92.73% and 78.94%, respectively, which shows that the rooting of fruit trees is difficult. The rooting rate of cuttings is closely related to the intensity of their own physiological activities [33].

2.2. Physiological Activities of Cuttings

Respiration, photosynthesis, and transpiration are the main physiological activities during tree cuttings, and the proper performance of the three kinds of physiological activities can effectively ensure the final cutting effect [34]. Cutting respiration is a process in which organic matter releases energy by oxidative decomposition under the action of a series of enzymes and hormones [35]. When the cells of the cuttings breathe aerobically, the accumulated organic matter is completely oxidized into CO₂ and H₂O [36]. Photosynthesis of cuttings is a biochemical process in which chlorophyll is irradiated by visible light [37]. CO₂ and H₂O are converted into organic matter through a light reaction, then O₂ is released [38]. These organic substances provide carbon sources for tree cuttings and roots, which are used for the growth, cell division, and tissue expansion of cuttings [39]. Transpiration is the main driving force of water circulation in cuttings [40], which enables cuttings to absorb water from the substrate and transport it to the tissues [41]. At the same time, three kinds of physiological activities will indirectly affect the temperature and humidity of the internal environment through the media of water and energy [42], as shown in Figure 3.

The characteristics of arbor plant cuttings include external characteristics and internal physiological activities. The external characteristics of cuttings will determine the strength of their physiological activity. In the cutting process of plants, since the cuttings have not yet formed roots and cannot effectively absorb water in the substrate, only one or two young leaves or one or two buds are retained to weaken the transpiration and respiration of the cuttings and maintain photosynthesis so as to ensure the balance of the physiological activities of the cuttings [43]. Physiological activities of cuttings are a series of chemical and physical changes carried out by tissue cells with the participation of plant hormones [44,45], which belongs to the common growth characteristics of cuttings. Therefore, it is necessary to summarize the functions of plant hormones in cuttings and the mechanism of environmental factors’ influence on plant hormones.

3. Characteristics of Plant Hormones in Cuttings

3.1. Types of Plant Hormones in Cuttings

Plant hormones play a crucial regulatory role in the growth and development of cuttings. They possess the ability to regulate cell division, elongation, and differentiation in cuttings, influencing the overall growth and development process [46,47]. During the rooting process of cuttings, prominent endogenous hormones that exert significant influence on them include auxins, abscisic acids, and cytokinins [48,49]. These plant hormones are primarily produced in the cambium layer, located between the phloem and xylem of the cutting [50]. In the cambium layer of the cutting, plant hormones generated during the active cell division process facilitate the cuttings’ rooting and growth [51].

The growth and development stage of cuttings is an intrinsic factor influencing the production of endogenous plant hormones [52,53]. Softwood cuttings and hardwood cuttings are different morphologies of cuttings at varying growth stages [54,55]. Compared with hardwood cuttings, softwood cuttings exhibit higher levels of auxins and cytokinins and lower levels of abscisic acid [56,57]. To enhance the rooting rate of cuttings, researchers apply exogenous hormones to cuttings before the cuttage [58,59,60]. Exogenous plant hormones can regulate the levels of endogenous plant hormones. The exogenous hormones commonly used for cuttings include indoleacetic acid (IAA), naphthaleneacetic acid (NAA), and indolebutyric acid (IBA). They belong to the auxin class of hormones, enabling the plant to promote the elongation and division of cutting cells [61,62]. The exogenous plant hormone 6-Benzylaminopurine (6-BA) is a synthetically produced cytokinin that can promote the differentiation and growth of cuttings. As a result, it facilitates the generation of adventitious roots [63,64]. Shi et al. treated linden cuttings with the exogenous hormone IBA, promoting the content of the endogenous hormone IAA to increase during the rooting induction period. Therefore, it facilitates a shift in the cuttings toward the physiological state favorable for rooting [65]. Wang et al. treated the cuttings of Fraxinus chinensis with 200 mg/L IBA. They found that, during the rooting process, the treated group exhibited a higher rate of increase in IAA content compared to the control group, thus demonstrating the regulatory effect of IBA on endogenous hormones [66]. Exogenous growth hormones significantly influence the nutrients, intracellular enzymes, chlorophyll, free amino acids, and other substances in the cuttings [67,68]. Zhu et al. [69] researched the cuttage of linden, treating the cuttings with IBA and water, respectively. The results indicated that the rooting rate of cuttings treated with IBA (75%) was significantly higher than those treated with water. Xu et al. [70] investigated the cuttage of Parthenocissus quinquefolia and found that the highest rooting rate was observed in those treated with 300 mol/L indole-3-acetic acid (IAA). Additionally, Li et al. [71] conducted a cuttage study on European spruce, discovering that different hormone types had varying effects on the rooting rate of cuttings, and higher concentrations of hormones were found to hinder rooting. Meanwhile, they found that the combined use of growth hormones resulted in better rooting effects.

3.2. Influence of Temperature and Humidity on Plant Hormone Functions

The temperature and humidity of the environment are external factors influencing the production and transport of plant hormones [72,73]. Within a certain range, an increase in temperature promotes the synthesis and transport of auxins in the cutting [74]. Auxins, in turn, stimulate the respiration of cutting cells, enhance cellular metabolic capacity, and provide energy and nutrients for the rooting of cuttings [75]. Furthermore, by adjusting the opening of stomata, auxins can influence gas exchange and water transpiration in the cuttings, thereby modulating the respiration and transpiration processes of cuttings [76]. Excessively high or low temperatures can lead to increased abscisic acid content within the cuttings, which sometimes may accelerate leaf abscission [77]. The number of leaves directly affects the intensity of photosynthesis and respiration of the cuttings [78,79]. Cytokinins promote cell division and tissue differentiation [80]. Within a certain temperature range, an increase in temperatures contributes to the synthesis of cytokinins [81]. Cytokinins affect the opening of stomata, thereby influencing transpiration efficiency. The dedifferentiation of the cutting cells is primarily controlled by auxins and cytokinins. When the auxin content exceeds that of cytokinins, it induces the dedifferentiation of tissue cells and the formation of root primordia [82].

Humidity also affects the synthesis and distribution of hormones within cuttings. A suitable humidity level helps maintain the water balance inside and outside the cuttings, thus preserving normal cell function and metabolic activity [83]. Conversely, high humidity may inhibit the formation of abscisic acids and cytokinins. In addition, both excessively high and low humidity can interfere with the distribution and transport of auxins [84]. The main mechanism through which humidity influences the levels of plant hormones is through the regulation of hormone synthesis and transport via the control of the plant stomatal opening and closing [85]. Excessive or insufficient air humidity can lead to stomatal closure, hindering the entry of carbon dioxide and the release of water vapor, thereby reducing photosynthetic efficiency [86]. Additionally, within a certain humidity range, an increase in humidity enhances the respiratory activity of plants [87]. However, if humidity is excessively high, it may lead to stomatal closure, hindering the entry of oxygen into the plant and thereby suppressing the respiratory activity [88]. Meanwhile, excessive humidity can also result in an excess of water within the plant, disturbing cellular metabolic processes and further affecting respiration [89]. The main way humidity influences transpiration is by affecting the opening and closing of stomata as well as vapor pressure differentials [90]. Excessive high atmospheric humidity reduces the vapor pressure differential between the inside and outside of the leaves, making it difficult for water vapor in the substomatal cavity to diffuse, consequently weakening transpiration [91].

The rooting process of cuttings involves the coordinated action of various plant hormones. Temperature and humidity affect the synthesis, transportation, and stability of plant hormones. Therefore, employing facility agriculture systems to maintain an appropriate temperature and humidity range is of great importance for promoting the rootage of cuttings and increasing survival rates.

4. Environmental Requirements for Cuttings

In a facility agriculture system, the temperature and humidity of the substrate and air can both influence hormones in cutting cells, which in turn affects the physiological activities of cuttings [92]. The environment required by cuttings is shown in Figure 4.

According to the presence or absence of substrates in the environment where the cuttings are planted, facility agriculture systems are divided into substrate environment-based and air environment-based systems. Due to the active physiological activities of plant cells at high ambient temperatures, nutrients will be transported to the high ambient temperature part in a concentrated manner, so the ambient temperature of the substrate should be higher than that of the air environment when cutting [93]. In this way, the cells at the cutting incision will obtain more nutrients, which is conducive to the formation of a callus. If the ambient air temperature is higher than the substrate ambient temperature, the buds and leaves of the cuttings will preferentially consume nutrients for respiration, resulting in a decrease in the rooting rate of the cuttings [94].

Table 2. Environmental parameters of cutting of apple rootstock shoots.

Growth Stage	Cuttage 0–12 Days				Cuttage 13–16 Days				Cuttage 17–30 Days
Time	5:00–19:00		19:00–5:00		5:00–19:00		19:00–5:00		5:00–19:00		19:00–5:00
Environment	Air	Substrate	Air	Substrate	Air	Substrate	Air	Substrate	Air	Substrate	Air	Substrate
Temperature (°C)	22	24	18	20	24	26	16	18	24	26	16	18

shows the environmental parameters of apple rootstock shoot cuttings, and the temperature of the substrate is higher than that of the air. The substrate is composed of vermiculite.

At the time of cutting, the relative humidity of the air and the moisture of the substrate are required to be 80–90% and 50–60%, respectively [95], which can form a film of water on the cutting surface [96]. Higher air humidity can reduce the evaporation of substrate water so that the substrate can maintain the appropriate humidity to reduce the water loss from the cutting cells to avoid the cuttings’ dehydration and an injury to seedlings [97]. In addition, substrate characteristics are also important factors affecting the rooting rate of arbor cuttings [98]. A substrate is used to provide growth support and nutrient supply for plant cuttings and seedlings. Therefore, when selecting and managing substrates, factors such as the demand of cuttings and environmental conditions should be considered [99]. Wang et al. took maple birch as the research object and verified the effect of the substrate type on cuttings through experiments [100]. Zhang viewed the substrate composition as the main factor affecting cutting propagation and proved that the substrate composition was also an important factor significantly affecting cuttings [101].

It can be seen from Figure 3 that there is a coupling relationship between temperature and humidity in the air environment and the substrate environment [102]. To study the relationship between the two, it is necessary to not only study the separate action mechanism of the influencing factors of air temperature and humidity and substrate temperature but also identify the interaction mode between the two. Therefore, it is of great significance to summarize the factors affecting the air environment, substrate environment temperature, and humidity to improve the environmental control effect of cuttings.

4.1. Factors Affecting the Ambient Temperature of Facility Agricultural Systems

The temperature in facility agriculture systems was mainly determined by heat conduction, heat convection, and heat radiation [103,104]. Conduction can transfer heat through direct collisions between molecules as well as the transfer of energy. Meanwhile, the rate of conduction is affected by the thermal conductivity, heat capacity, temperature difference, and distance between objects [105]. Convection is the process by which heat is transferred through the flow of a gas or liquid, where a fluid that expands by heat or shrinks by cold transfers heat from one area to another [106]. The rate of convective heat transfer depends on the nature of the fluid, the velocity of the flow, the temperature difference, and the shape and structure of the object surface [107]. Radiation is the process of transferring heat through electromagnetic waves. The rate of thermal radiation depends on the temperature of the object, the emissivity of the surface, and the distance between the objects [108]. Heat transfer in the internal environment of facility agriculture systems is usually a combination of the above three methods [109].

4.1.1. Factors Affecting Air Ambient Temperature

In the air environment of a facility agriculture system, the heat transfer medium is mainly gas molecules and water vapor, whose content and density determine the heat transfer effect. The ambient temperature of the air mainly heats the air near the heat source by means of heat radiation and heat conduction. The heated air then transfers heat to the cooler part in the form of heat convection. Therefore, flow conditions will affect the change in heat transfer path and the uniformity of heat distribution [110].

For the case where the heat transfer medium is gas, the flow can be divided into natural convection and forced convection. Natural convection is caused by the difference in gas density under the action of gravity in the absence of external driving forces [111,112]. Natural convection in the process of heat transfer in air environment can form convective circulation [113]. When the gas is heated and the temperature difference is generated, the surrounding air density changes. The gas in the area with lower density will rise, while the gas in the area with higher density will sink, thus forming convective circulation [114]. Forced convection is generated by an external driving force [115], including a mechanical device (such as a fan) or other form of external force, which often produces a stronger convection effect and a higher heat transfer rate than natural convection [116]. In the air environment, convection heat transfer is usually faster than the conduction heat transfer rate, and the flowing medium can form a new contact surface of cold and hot media through movement, thus improving the heat transfer effect [117].

The air environment of the facility agricultural system can be heated by means of sunlight, hot blast stove heating, a fin radiator, etc., and it can be cooled by means of a water curtain, a negative pressure fan, and the location of tuyere and ventilation forms together with air conditioning [118,119] (Figure 5).

4.1.2. Factors Affecting Substrate Ambient Temperature

The heat transfer of the substrate is mainly through heat conduction and heat radiation. Light intensity has a certain influence on the temperature of the substrate [120] because the absorption and reflection of light by the substrate will cause itself to be heated [121]. The color of the substrate affects the effect of heat radiation energy transfer because the color of different types of substrates has a different ability to absorb and release heat. For example, the dark substrate absorbs more heat more easily, while the light substrate has a strong reflection ability and does not easily rise in temperature through thermal radiation [122], as shown in Figure 6. The physical properties of heat transfer in the substrate have different influences on the heat transfer, including thermal conductivity, density, heat capacity, and thermal diffusivity [123,124]. The parameters of heat transfer characteristics of the substrate commonly used in plant cuttings are shown in

Table 3. Comparison of substrate heat transfer characteristics [127,128,129].

Substrate Material	Heat Conductivity (W/(m·K))	Density (g/cm3)	Calorific Capacity (J/(g·K))	Thermal Diffusivity (mm²/s)
River Sand	0.2–1.5	1.4–2.7	0.8–1.2	0.1–1.0
Perlite	0.06–0.2	0.8–2.9	0.8–1.2	0.1–1.0
Turfy Soil	0.03–0.4	0.1–0.3	1.0–1.5	0.1–1.0
Vermiculite	0.04–0.07	2.2–2.8	0.8–1.2	0.1–1.0
Rice Hull	0.03–0.1	0.1–0.2	1.4–1.8	0.1–1.0
Redheart Soil	0.2–1.5	1.2–1.8	0.8–1.5	0.1–1.0
Coconut Bran	0.03–0.1	0.1–0.2	1.5–2.5	0.1–1.0

. There is water between the pores of the substrate, and the temperature of water will affect the heat transfer of the substrate; the water retention of the substrate is also a factor affecting the heat transfer characteristics of the substrate [125,126]. Therefore, the temperature of the substrate was affected by irrigation frequency, water quantity, and water temperature. Since irrigation also directly changes substrate humidity, irrigation methods are described in the section on substrate humidity control methods.

As shown in Figure 7, the temperature in the substrate environment can be controlled by means of buried pipes, geothermal wires, and thermostatic master disc water baths. The buried pipe sends water into the pipe to change the temperature of the substrate via heat conduction [130], and the geothermal wire uses the resistance wire to heat the substrate [131]. The water bath temperature control master designed by Mi is used for cutting and breeding apple rootstock, which can control the temperature of the substrate by controlling the circulation of water and the heat exchange with the substrate [132].

4.2. Factors Affecting Environmental Humidity in Facility Agricultural Systems

4.2.1. Factors Affecting the Humidity of the Air Environment

There is a close relationship between humidity and temperature in the air environment. As shown in Figure 8, they jointly determine the moisture content in the air [133]. The increase in temperature causes the saturated water vapor pressure in the air to increase so that the air can hold more water [134]. The moisture contained in the air is generally called relative humidity, which is expressed as a percentage of the ratio between the actual water vapor pressure and the saturated water vapor pressure. The increase in temperature will make the saturated water vapor pressure in the air rise faster than the actual water vapor pressure, resulting in the decrease in relative humidity [135]. Under specific humidity conditions, the corresponding temperature when the water vapor in the air reaches saturation is called the dew point temperature [136]. When air is cooled to the dew point temperature, condensation occurs, resulting in a decrease in relative humidity [137,138]. The dew point temperature is related to relative humidity and temperature, and with the increase in relative humidity or the decrease in temperature, the dew point temperature will decrease [139]. Therefore, changes in temperature affect the relative humidity and dew point temperature, thereby changing the moisture content in the air [140].

The water vapor condensation usually starts with a condensation nucleus. A condensation nucleus in facility agriculture systems includes tiny solid or liquid particles, and when saturated vapor pressure and temperature in the air remain constant, the evaporation of a condensation nucleus reaches dynamic equilibrium [141]. In addition, condensation is also affected by air flow, which affects water vapor condensation mainly through auxiliary mass transfer to increase the contact opportunity and remove the saturated layer [142]. The diffusion between water vapor and air is accelerated during air flow, which helps to bring water vapor molecules around the condensation nucleus. Water vapor condenses into small droplets on the condensation nucleus to form a saturated layer, which prevents new water vapor from condensing. The moving air removes the saturated layer around the condensed water, enabling the condensation to occur. However, too strong or too fast of an air flow may cause water vapor to disperse and dilute, thus slowing the condensation [143].

The water vapor in the air environment can be supplemented by the atomizing device, and the spray can be divided into different forms according to the size of the droplets. It mainly includes aerosol spray, mist spray, spray drift, and drift spray [144,145]. The droplets formed by an atomizing spray are subtler and can stay suspended in the air for a long time. The droplets formed by a mist spray are very small, like tiny water droplets suspended in the air. The droplets form large droplets that fall quickly to the ground. The droplets formed by a drift spray are larger and need to be influenced by wind to travel long distances [146]. The parameters of different spray forms are shown in

Table 4. Parameters of different spray forms [144,145,146].

Spray Form	Droplet Size (μm)	Suspension Time
Aerosol Spray	1–10	Long
Mist Spray	10–100	Comparatively Long
Spray Drift	100–1000	Comparatively Short
Drift Spray	1000–100,000	Short

.

To maintain the humidity of cuttings in the air environment, it is necessary to not only choose the appropriate atomization equipment but also know about the structure of the cuttings’ surface because the surface structure of cuttings has a significant influence on the dynamic behavior of droplet wall impact [147]. The topological morphology of the cuttings’ surface determines its wetting characteristics [148], which are usually divided into hydrophobic, superhydrophobic, hydrophilic, and superhydrophilic characteristics, and the degree of infiltration directly determines the amount of droplets on the cuttings’ surface. The topological morphology of cuttings includes surface pores [149] and irregular attachments (hairs, spines, etc.) [150] (Figure 9 and Figure 10).

$$\left\{\begin{array}{l}{\gamma }_{\mathrm{SV}}-{\gamma }_{\mathrm{SL}}={\gamma }_{\mathrm{LV}}\cos \theta \\ \cos {\theta }_w=r\cos \theta \\ \cos {\theta }_c=f_s\cos {\theta }_s+f_{{}_V}+\cos {\theta }_V\end{array}\right.$$

(2)

In the Equation (2), r represents the ratio between the apparent solid contact area and the intrinsic solid contact area, r ≥ 1, $f_s$ + $f_v$ = 1.

The dynamic morphological changes of a single droplet impacting a solid wall can be divided into six categories [153], as shown in Figure 11.

In the critical stage of the impact process, forced energy and mass are conserved [154]. Roisman et al. found that when a single droplet has a retreat after impact and diffusion, the process is determined by the interaction between the free edge and the internal liquid layer and the force exerted by the wall on the edge [155]. Kim and Chun supplemented the energy balance formula [156] and approximated the shape of the water drop as a cylinder, which showed that it could describe the spreading and droplet retreat behavior well. Xu et al. used the spring-mass-damping equation to describe the recoil stage process and determined its coefficient semi-empirically with actual physical parameters [157]. Okumura demonstrated by scale that a droplet can behave as a spring mass system under the limit of small deformation [158]. The modeling of the rebound and crushing stage mostly adopts the traditional energy balance principle [159]. The researchers used experimental fitting to produce a probabilistic method for predicting the adhesion, rebound, and splash [160]. The criterion for a rebound was first proposed by Mao et al. to determine whether the residual energy after impact could cause a droplet to rebound, and they predicted the rebound behavior of a droplet impacting a horizontal surface from directly above through a series of energy balances [161]. Dorr et al. extended this model by adding multi-angle impact scenarios [162] and predicted the occurrence of a rebound by calculating the “Excess Rebound Energy” (EERE), which occurs when EERE > 0. In fragmentation stage modeling, Mundo et al. proposed the basis for determining whether the droplet’s kinetic energy can overcome the capillary effect and lead to droplet fragmentation [163], as shown in Equation (3).

$$K=\mathrm{Oh}{(\mathrm{Re})}^{1.25}={\mathrm{We}}^{0.5}{\mathrm{Re}}^{0.25}>K_{\mathrm{crit}}$$

(3)

The K describes the factors related to droplet properties, and K_crit characterizes the surface characteristics of the impact wall, including roughness and wettability. When K > K_crit, crushing occurs; when K < K_crit is applied, the droplet will adhere or rebound, and the specific behavior of the droplet can be further identified by combining the residual energy formula in the Dorr model.

Therefore, in order to meet the requirements of water film formation on the cuttings’ surface, it is necessary to clarify the characteristics of the cuttings as the impacting surface of droplets. The wetness of the cuttings’ surface is not only related to the moisture content in the air environment but also the amount of water absorbed via transpiration from the substrate.

4.2.2. Factors Affecting the Ambient Humidity of the Substrate

The substrate consists of one or more granular materials, belonging to a porous medium. Relative humidity is an important factor affecting the moisture of the substrate [164]. When the relative humidity of the air environment is high, the substrate more easily absorbs the moisture in the air, resulting in an increase in its own humidity [165]. When the relative humidity is low, the water in the substrate more easily evaporates [166,167]. The water retention capacity of the substrate was determined by the type of substrate and the drainage capacity, while different types of substrates had different water retention capacities. The composition of the substrate determines the overall pore structure and affects its ability to adsorb and release water, which is reflected in the strength of the substrate’s drainage capacity [168]. Higher temperatures increase the rate of evaporation of water in the substrate, resulting in lower substrate humidity [169,170]. The facility agriculture system can increase the air flow rate at the surface of the substrate and accelerate water evaporation through the capillary transport up the pores, thus changing the substrate humidity [171]. The irrigation frequency and single irrigation amount directly affect the humidity of the substrate [172,173] (

Table 5. Parameter comparison of substrate particle size characteristic.

Substrate Material	Size	Particle Size (mm)	Water-Holding Capacity	Water Discharge Capacity	Reference
River Sand	Gravel	>2 mm	Relatively Weak	Relatively Strong	[174,175,176,177]
	Coarse Sand	0.25–2
	Medium Sand	0.12–0.25
	Fine Sand	0.06–0.125
	Very Sine Sand	0.03–0.06
Perlite	Coarse Particle	2.12–5.38	Relatively Good	Relatively Weak	[178,179]
	Medium Particle	0.51–2.01
	Fine Particle	0.13–0.52
Peat	Coarse Particle	2.21–5.32	Good	Weak	[180,181]
	Medium Particle	0.53–2.23
	Fine Particle	0.14–0.48
Vermiculite	Coarse Particle	2.09–5.22	Relatively Good	Relatively Weak	[182,183,184]
	Medium Particle	0.52–2.01
	Fine Particle	0.13–0.53
Peanut Shell	Coarse Particle	2.17–5.32	Good	Weak	[185,186]
	Medium Particle	0.48–2.01
	Fine Particle	0.11–0.51
Coconut Chaff	Coarse Particle	2.24–5.13	Good	Weak	[187,188]
	Medium Particle	0.53–2.02
	Fine Particle	0.12–0.52

).

The methods of substrate humidification in the facility agriculture system include surface drip irrigation, insert drip irrigation, and spraying [189,190,191], as shown in Figure 12. Irrigation methods will lead to different volumes of substrate-wetted bodies [192], as shown in Figure 13.

5. Construction Methods of Environmental Model in Facility Agriculture Systems

The environmental model of a facility agriculture system mainly includes the mechanism model and data-driven model [193]. The mechanism model is a dynamic model based on the principles of energy conservation and mass conservation [194]. The data model, also known as the black box model, is based on modern computational theory and builds a process model according to the characteristic information provided by the data.

5.1. Environmental Mechanism Model of Facility Agriculture Systems

Businger [195] divided the interior of the greenhouse into four layers based on the balance between energy and mass, and his research method became the basis for the mechanism modeling of plant factories. Pasgianos [196] used the computational fluid dynamic (CFD) theory and method to analyze and study the air exchange rate of natural light-type plant factories by taking the temperature difference, outdoor wind speed, and skylight opening angle as parameters. Boulard and Wang used CFD to study climate distribution and crop transpiration heterogeneity inside the greenhouse and studied the mechanism changes of temperature and humidity from the perspectives of plant physiology and thermodynamics [197]. Liu used computational fluid dynamics to calculate the flow field as well as temperature field and studied an adjoint method to achieve optimal ventilation design in a closed environment [198]. Xu [199] proposed a CFD-based cooling environment optimization design method for a greenhouse wet curtain–fan system, which improved the environmental performance of greenhouse cooling in summer. Saberian developed a computational fluid dynamic model of a full-scale gabled greenhouse covered with translucent materials to simulate the effect of the solar dynamic heat load on the greenhouse microclimate so as to predict microclimate dynamics in 2017 and 2018 [200]. The microclimate model constructed by Singh et al. describes the transfer process of energy and matter, and this model can predict the temperature of air, crops, and cultivation substrates under natural ventilation conditions [201]. Li built a computational fluid dynamic model for greenhouse buildings to achieve the multi-objective and high-efficiency optimization of greenhouse environmental factors, in which the thermal environment component of the greenhouse was the main research object [202]. Boulard et al. established a fluid dynamic model to predict the distribution of temperature and water vapor in the glass greenhouse, optimize the micrometeorological conditions in the greenhouse, and provide a theoretical basis for energy saving [203]. Zhang et al. used a CFD numerical simulation to analyze the 20 °C isosurface of the rear wall of the greenhouse for heat storage and determined that the reasonable length of the hot air storage duct is 20 m [204]. Above such models can calculate the distribution of temperature and humidity (spatial characteristics) in greenhouses (

Table 6. Studies on environmental mechanism models.

Model	Construction Principle	Instrument	Study Subjects	Key Influence Factor	Model Type	Evaluation Index	References
Air Exchange Rate Model	Conservation of Energy, Conservation of Mass	Matlab	Temperature, Humidity	Ventilation Flow Rate, Evaporation Rate, Atomized Water Volume, Air Heat Capacity	Nonlinearity	Average Error (MAE)	Pasgianos [196]
Plastic Tunnel Greenhouse Climate and Crop Heat Exchange	Conservation of Energy, Conservation of Mass	CFD	Temperature, Humidity	Solar Radiation Intensity, Transpiration, Ventilation Speed	Nonlinearity	Average Error (MAE)	Boulard [197]
Climate Distribution Model in Greenhouse	Conservation of Energy, Conservation of Mass	CFD	Temperature	Ventilation Flow, Intake Temperature, Intake Speed	Nonlinearity	Average Error (MAE)	Liu [198]
Summer Greenhouse Cooling Simulation Model	Conservation of Energy, Conservation of Momentum, Conservation of Mass	CFD	Temperature	Solar Radiation Intensity, Crop Physiological Activity, Wind Speed, Wet Curtain Area, Greenhouse Length	Nonlinearity	Average Error (MAE), Relative error (MRE), Maximum Absolute Error (MaxE)	Xu [199]
Microclimate Model of Gable Greenhouse	Conservation of Energy, Conservation of Mass	CFD	Temperature, Humidity	Solar Radiation Intensity, Ventilation Wind Speed, Ventilation Direction	Nonlinearity	Root-Mean-Square Error (RMSE)	Saberian [200]
Greenhouse Temperature Prediction Model Under Natural Ventilation	Conservation of Energy, Conservation of Momentum, Conservation of Mass	Matlab	Temperature	Crop Leaf Surface Temperature, Indoor Air Temperature, Soil Temperature	Nonlinearity	Determination Coefficient (R2), Standard Deviation ($\sigma$) Root-Mean-Square Error (RMSE) Model Efficiency (${\eta }_{eff}$)	Singh [201]
Spatial and Temporal Variation Characteristic Factor Optimization Model of Greenhouse Environment	Conservation of Energy, Conservation of Mass	CFD	Temperature, Humidity, Energy Consumption	Soil Temperature, Soil Density, Air Density, Indoor Air Temperature, Greenhouse Roof Temperature, Wind Speed	Nonlinearity	Maximum Relative Error (MaxRE), Average Relative Error (ARE), Root-Mean-Square Error (RMSE)	Li [202]
Temperature and Water Vapor Distribution Model in Glass Greenhouse	Conservation of Energy, Conservation of Mass	CFD	Temperature, Humidity	Crop Leaf Area, Solar Radiation Intensity, Plant Cover Temperature, Photosynthesis, Respiration, Transpiration Heat, Water Vapor Volume	Nonlinearity	Root-Mean-Square Error (RMSE)	Boulard [203]
Greenhouse Thermal Storage Rear Wall Model	Conservation of Energy, Conservation of Mass	CFD	Temperature	Wall Temperature, Hot Air Duct Temperature, Length of Hot Air Duct	Nonlinearity	Absolute Error(AE), Average Error (MAE), Average Relative Error (ARE), Maximum Relative Error (MaxRE)	Zhang [204]

and Figure 14).

5.2. Environmental Data-Driven Model of Facility Agriculture Systems

Data-driven modeling can be divided into the statistical regression method and shallow neural network method [205,206]. Yu et al. proposed a temperature prediction model based on the least squares support vector machine (LSSVM). Improved particle swarm optimization (IPSO) was adopted to optimize the parameters of the LSSVM model, and the short-term prediction effect of the future was better [207]. Since the greenhouse system has the characteristics of slow time-varying and the time-series data of the greenhouse environment are characterized by specific trends and cycles, the greenhouse time-series modeling can be used to study the specific law of data change. Ren et al. designed and developed a greenhouse environment monitoring and temperature prediction system based on the differential time-series model by combining the Internet of Things, cloud services, and wechat platform [208]. Du et al. proposed a nonlinear autoregressive dynamic neural network model for solar greenhouse temperature predictions [209]. Jung et al. built an artificial neural network, nonlinear autoregressive model, and long short-term memory model (LSTM) to predict air temperature and humidity [210]. Guillen-Navarro MA et al. used LSTM and 4-month temperature data to predict in advance whether there would be an extremely low temperature in the greenhouse, and the prediction error was less than 0.8 °C [211]. Hu et al. proposed a two-dimensional substrate input network based on 1D CNN-GRU structural characteristics and time-step length for temperature prediction, providing an important basis for accurate and efficient temperature control of facility agriculture systems [212]. In the study of short-term time series under nonlinear and non-stationary scenarios, some scholars proposed to combine the autoregressive integrated moving average model (ARIMA) [213,214], genetic algorithm (GA), and support vector regression (SVR) [215], emphasizing the capture of different model features within the data to improve the prediction performance of a single model so as to obtain temperature prediction values more in line with the actual situation and improve the accuracy and effectiveness of temperature prediction. Zou constructed a convex bidirectional extreme value learning machine (CB-ELM) to establish a solar greenhouse temperature and humidity prediction model, reducing potential economic losses [216]. In order to solve the hysteresis problem in solar greenhouse temperature regulation, Li proposed a solar greenhouse temperature prediction model based on a BP neural network [217].

The data-driven model trains a large number of high-quality data based on statistical laws, improves adaptively according to the data, and continuously optimizes its own parameters and structure so as to improves the prediction accuracy to cope with various complex and changeable environmental conditions. The data-driven model has advantages in dealing with nonlinear relationships and can accurately capture complex nonlinear dynamics in the environment of a facility agriculture system and mine hidden information and patterns in the data. However, current studies on temperature and humidity prediction of facility agricultural systems based on deep learning mostly focus on a single temporal network [218], unable to mine deep and effective information among the environmental data, and the maximum prediction time of the model is only 1 h. Further studies are needed to meet the needs for the accuracy, efficiency, and duration of the model [219] (

Table 7. Research on environmental data-driven model.

Model	Type	Instrument	Study Subjects	Key Influence Factor	Model Type	Evaluation Index	References
Temperature Prediction Model	Statistical Regression Method	Matlab, Least Squares Support Vector Machine (LSSVM), Intelligent Particle Swarm Optimization (IPSO)	Temperature	Air Temperature, Air humidity, Soil Temperature, Soil moisture, Solar Radiation Intensity, Outdoor Temperature	Nonlinearity	Mean Absolute Error (MAE), Mean Percentage Error (MAPE), Mean Square Error (MSE)	Yu [207]
Temperature Prediction System	Neural Network	Time Sequence Forecast Model (ARMA)	Temperature	Air Temperature, Air humidity, Light Intensity	Nonlinearity	Maximum Absolute Error (MaxAE), Maximum Relative Error (MaxRE), Mean Relative Error (MRE)	Ren [208]
Solar Greenhouse Temperature Prediction Model	Neural Network	Model Predictive Control (MPC), Nonlinear Autoregressive Exogenous Model (NRAX)	Temperature	Wind Speed of Fan, Wet Curtain State, Solar Radiation Intensity, Air Humidity	Nonlinearity	Mean error (ME)	Du [209]
Velon Greenhouse Climate Prediction Model	Neural Network	Artificial Neural Network (ANN), Nonlinear Autoregressive Exogenous Model (NRAX), Recurrent Neural Network (RNN), Long Short-Term Memory Network (LSTM)	Temperature, Humidity	Air humidity, air Temperature, CO2 Concentration	Nonlinearity	Mean error (ME), Root-Mean-Square Error Prediction (RMSEP), Standard Error Prediction (SEP), Determination Coefficient (R2)	Jung [210]
Agricultural Low Temperature Prediction Model	Neural Network	Long short-term memory network (LSTM)	Temperature	Air Temperature, Air humidity, Wind Speed	Nonlinearity	Root-Mean-Square Error (RMSE), Determination Coefficient (R2), Mean Absolute Error (MAE), Pearson’s Correlation Coefficient (PCC), Error Percentage	Guillén-Navarro M Á [211]
Solar Greenhouse Temperature Prediction Model	Neural Network	One-Dimensional Neural Network, Gated Cycle Unit (GRU), Convolutional Neural Network (CNN)	Temperature	Light Intensity, Indoor Soil Temperature, Outdoor Soil Temperature, Outdoor Air Temperature, Outdoor soil Moisture Content, Outdoor CO2 Concentration	Nonlinearity	Determination Coefficient (R2), Root-Mean-Square Error (RMSE), Mean Absolute Error (MAE), Maximum Absolute Error (MaxAE), Average Relative Error (ARE)	Hu [212]
Prediction of Greenhouse Temperature of Edible Fungi	Statistical Regression Method	Moving Average (MA), Autoregressive Integrated Moving Average Model (ARIMA), Genetic Algorithm (GA), Support Vector Regression (SVR)	Temperature	Moving Average Window Length, Indoor Temperature at 3 o’ clock	Nonlinearity	Mean Relative Error (MAE), Mean Square Error (MSE), Mean Absolute Percentage Error (MAPE)	Tian [215]
Solar Greenhouse Temperature and Humidity Prediction Model	Statistical Regression Method	Convex Bidirectional Extremum Learning Machine (CB-ELM)	Temperature, Humidity	Solar Radiation Intensity, Wind speed, Outdoor Temperature, Outdoor Humidity	Nonlinearity	Root-Mean-Square Error (RMSE), Model Validity	Zou [216]
Solar Greenhouse Temperature Prediction Model	Neural Network	Backpropagation Neural Network (BP)	Temperature	Indoor Temperature, Indoor Humidity, Indoor Light Intensity	Nonlinearity	Mean Absolute Error (MAE), Mean Relative Error (MRE), Maximum Absolute Error (MaxAE)	Zhao [217]

and Figure 15).

Two types of model pairs are shown in

Table 8. Comparison between mechanism model and data-driven model.

Model	Mechanism Model	Data-Driven Model
Common point	The main research objects are ambient temperature and humidity, among which temperature accounts for more. All models are nonlinear models. They are predictive.
Advantages	A process-oriented visual description of the space–time change of matter	Fast induction of input/output relationships for high dimensional data
Disadvantages	The boundary conditions need to be tested in practice	The quality of training data is high and the amount of feature data is large
Individualized improvement plan	The boundary conditions that are difficult to measure are derived by using deep learning methods	The mechanism model is used to screen the influential factors and reduce the dimension of the input data
Common improvement plan	The content and distribution of plant hormones were used as important parameters to analyze the intensity of plant physiological activities. They should be important parameters for constructing environmental models.

.

6. Discussion, Conclusions, and Prospects

6.1. Discussion

(1) The survival rate of cuttings is mainly determined by both inherent characteristics of the cuttings themselves and external temperature and humidity. The inherent characteristics of cuttings include their external appearance and the physiological activities within cuttings under the influence of hormones. As physiological commonalities in the growth and development of cuttings, environmental temperature and humidity influence cuttings’ physiological activities by affecting the production and distribution of hormones. At the same time, the physiological activities of cuttings also impact environmental temperature and humidity in the form of water and energy. The characteristics of cuttings’ roots serve as the evaluation criteria for the effectiveness of cuttage. However, depending solely on root characteristics as the standard for evaluating cuttage effectiveness is insufficient to demonstrate cuttings’ health conditions during their growth process. Therefore, it is necessary to supplement the evaluation criteria further. For example, before the rootage of cuttings, the dynamic changes in the physiological state of tissue cells can be incorporated as the process evaluation criteria.

(2) In facility agriculture systems, the physiological activities of cuttings are maintained in balance by controlling the internal temperature and humidity of the agriculture system. The internal environment in facility agriculture systems is categorized into air and substrate environments. The heat transfer in the internal environment is the combined effects of heat conduction, heat convection, and thermal radiation. The humidity in both the air and substrate environments is determined by the moisture content, including the transformation and movement between gaseous and liquid water. The factors influencing the temperature and humidity in the internal environment are intricate, with coupling relationships among them. Consequently, when researching the construction of temperature and humidity models for arbor tree cuttings in facility agriculture systems, it is essential to explore the influence weights of various factors on temperature and humidity. Decoupling methods can be adopted to clarify the influence weights of each factor, and those with significant influence weights should be selected as parameters for building the environmental model.

(3) The environmental model in facility agriculture systems comprises mechanistic models and data-driven models, and the main research objects are temperature and humidity. Both types of models have their strengths in predicting temperature and humidity. Mechanistic models, predominantly employing computational fluid dynamics (CFD) as a key research tool, can vividly depict the spatiotemporal distribution of temperature and humidity, along with the energy changes induced by the physiological activities of cuttings. Data-driven models can simultaneously handle high-dimensional data, leveraging its powerful computational capabilities to predict the trends in temperature and humidity rapidly. Considering the large volume of physiological model data of cuttings themselves, coupled with numerous factors in facility agriculture systems affecting temperature and humidity as well as the coupling effects among these factors, it is advantageous to integrate mechanistic models with data-driven models when constructing environmental models. Mechanistic models can be employed to screen impact factors, consequently reducing the data dimensions for data-driven models and enhancing their convergence speed.

(4) Both the environmental mechanistic models and data-driven models can be constructed in Matlab, which is expected to remain an essential tool for building these two types of models in the future. However, with the ongoing progress in deep learning, there will be a gradual improvement in the quality of collected feature data, and the quantity will also become substantial. Therefore, deep learning neural networks built in the Python language might offer more advantages in constructing environmental models. The purpose of constructing environmental models is to better observe the distribution of temperature and humidity within facility agriculture systems. This provides a basis for the dynamic adjustment of environmental parameters in the control system of the internal environment, enhancing the comfort of cuttings in facility agriculture systems, ultimately improving the rooting and survival rates of the cuttings.

6.2. Conclusions

Rationally controlling the temperature and humidity is the key to improving the rooting rate of arbor cuttings. The temperature and humidity environmental model can provide a basis for regulating temperature and humidity in the control system. Therefore, we summarized the relationship between the characteristics of arbor cuttings, plant hormones’ characteristics, and environmental temperature and humidity. We combined the ways of constructing environmental models for facility agriculture systems to conduct research on the construction methods of temperature and humidity environmental models in arbor tree cuttage. The main research contents are as follows.

(1) The external traits of arbor cuttings include the type, the location part, the length of cuttings, the number of leaves and buds on the cutting branches, and the characteristics of the roots. The internal physiological activities of cuttings include photosynthesis, transpiration, and respiration. The external traits of cuttings influence the intensity of their physiological activities, while the physiological activities mutually influence each other in the forms of water and energy with environmental temperature and humidity. The external traits and internal physiological activities collectively constitute the characteristics of cuttings. Therefore, when constructing a temperature and humidity environmental model for arbor tree cuttings, monitoring the inherent characteristics of cuttings can estimate their physiological activity intensity.

(2) Plant hormones regulate the growth and development of cuttings by influencing their inherent physiological activities. Plant hormones are categorized into endogenous hormones and exogenous hormones. The endogenous hormones significantly impacting cuttings primarily comprise auxins, abscisic acids, and cytokinins. On the other hand, exogenous hormones applied before cuttage mainly include indoleacetic acid (IAA), naphthaleneacetic acid (NAA), indolebutyric acid (IBA), and 6-benzylaminopurine (6-BA). All endogenous hormones and exogenous hormones can regulate the division, elongation, and differentiation of cutting cells. In addition, exogenous plant hormones can modulate the levels of endogenous plant hormones. In summary, the content of and changes in plant hormones in cuttings influence the effectiveness of their growth and development. Therefore, by controlling the levels of plant hormones in cuttings, it is possible to regulate the intensity of the cuttings’ inherent physiological activities during the rooting process.

(3) Environmental temperature and humidity are the primary external factors influencing the generation and transport of plant hormones. In facility agriculture systems, the environment is classified into substrate and air environments based on the presence or absence of a substrate. Both environments encompass temperature and humidity parameters. Specifically, factors influencing the temperature of the air environment include the density and flow state of gas molecules as well as water vapor. The temperature of the air environment can be adjusted by leveraging corresponding facilities based on the principles of natural convection and forced convection. Similarly, factors influencing the temperature of the substrate environment include the color, heat transfer characteristics, and water retention capacity of the substrate. The temperature of the substrate environment can be regulated by utilizing facilities such as buried pipes, geothermal wires, and thermostatic master discs based on the principles of heat conduction and thermal radiation. Furthermore, factors influencing the humidity of the air environment mainly include the temperature of the air environment, the air velocity, and the role of condensation nuclei. The moist state of the cutting’s epidermis is primarily associated with the humidity of the air environment and the characteristics of the epidermis. Moisture in the air environment is supplemented by atomizing devices. In addition, the humidity of the substrate environment is influenced by the characteristics of the substrate and the relative humidity of the air environment. Methods such as surface drip irrigation, inserted drip irrigation, and sprinkling can alter the volume of substrate-wetted bodies. Understanding the factors influencing environmental temperature and humidity as well as the mechanisms by which temperature and humidity affect the levels of plant hormones in cuttings is a method for regulating the intensity of physiological activities in cuttings.

(4) Environmental models are constructed from the factors affecting environmental temperature and humidity in facility agricultural systems. They can describe changes in temperature and humidity. According to different modeling methods, it can be divided into a mechanism model and data-driven model. The building principles of the mechanism models include the law of energy conservation, the law of momentum conservation, and the law of mass conservation. The data-driven models are mainly constructed based on statistical regression and neural network. When constructing the temperature and humidity environment model of cuttings in the facility agriculture system, it is necessary to comprehensively consider the characteristics of cuttings themselves and the factors affecting the environmental temperature and humidity of the facility agriculture system. By changing the environmental temperature and humidity, the production, transport, and distribution of plant hormones in the cuttings are affected so as to control the physiological activity intensity of cuttings and improve the adaptability of the environmental model. Combining the visual description characteristics of the mechanism model and the ability of data-driven model to process high-dimensional data, the temperature and humidity environment model of arbor cuttings can be effectively constructed. The choice of the model depends on the temperature and humidity requirements of cuttings, the available data, the environmental conditions, and the cost of obtaining effective data.

6.3. Prospect

(1) The key to constructing environmental models for cuttings lies in the effective collection of data. The growth and development process of cuttings exhibit characteristics such as diverse feature types, large data volumes, and slow, continuous changes. Therefore, machine vision technology can be employed to monitor the growth conditions of cuttings; image segmentation techniques can be adopted to extract the morphology and changes in roots, stems, leaves, and buds of cuttings; imaging devices can be utilized to monitor the temperature distribution and moisture level on the surface of cuttings, capturing the opening and closing of stomata; and multispectral technology can be leveraged to identify and analyze the internal features of cuttings, including changes in moisture content and hormone quantities. Through continuous image data, the physiological activities of cuttings can be indirectly observed and analyzed.

(2) As fruit trees often exhibit a lower rooting rate in cuttings, further research on their cutting techniques is needed within facility agriculture systems, including studying the inherent physiological characteristics of fruit tree cuttings and their specific requirements for environmental temperature and humidity. Given that cuttings have no roots when initially inserted into the substrate and the intensity of their physiological activities differs from that of rooted plants, it is essential to categorize the rooting process of cuttings into distinct growth stages when constructing physiological models. These stages include the rootless period, callus tissue formation period, and adventitious root formation period.

(3) When designing the temperature and humidity environmental model for arbor tree cuttings, it should adhere to the principle of maintaining the balance of the physiological activities of cuttings. Environmental parameters of temperature and humidity should experience dynamic changes to facilitate the synthesis, transport, and distribution of hormones, thus maintaining a reasonable ratio of plant hormones within the cuttings and meeting the physiological commonality requirements for the growth and development of cuttings. Additionally, other factors affecting cuttings’ physiological activities, including enzymes and other environmental elements, should also be considered. Since the growth process of cuttings and the environmental temperature and humidity are sequential data characterized by continuous changes at time nodes, it is feasible to combine the growth-state model of cuttings with the environmental model. In the future, the physiological model of cuttings, the environmental model, and the performance of the control system can be integrated into a unified environmental identification model. This would facilitate the growth control of cuttings, thereby improving the universality and transferability of the temperature and humidity environment model.