Research on the Impact of Meteorological Disaster Shocks and Risk Perception on Farmers’ Cooperative Pest and Disease Control Behavior

Abstract

Climate change, primarily characterized by rising global temperatures, has led to a continuous expansion in the area affected by pests and diseases. This poses a significant threat to national agricultural production and directly jeopardizes food security. Cooperative pest and disease control behavior represents a novel approach to pest and disease management and is an important measure for mitigating agricultural production risks. This study employs Probit and IVProbit models to empirically examine the impact of meteorological disaster shocks on cooperative pest and disease control behavior. The results show that both meteorological disaster shocks and their frequency significantly and positively influence cooperative pest and disease control behavior. Moreover, the perception of risk losses plays a mediating role in this relationship. The impact of meteorological disaster shocks on farmers’ cooperative pest and disease control behavior is heterogeneous. Specifically, these shocks and their frequency significantly influence the choice of cooperative pest and disease control behavior among farmers with junior high school education or above and large-scale farmers, while they have no significant impact on farmers with primary school education or below and small-scale farmers.

1. Introduction

Against the backdrop of global climate change, agriculture, an industry highly susceptible to climate influences, is confronted with unprecedented challenges. The frequent occurrence of meteorological disasters, such as torrential rains, droughts, and hurricanes, not only poses a threat to the growth of crops but also seriously endangers global food security [1,2,3]. According to the data from the Food and Agriculture Organization of the United Nations (FAO), in recent years, the extent of crop yield reduction caused by meteorological disasters has been increasing year by year, dealing a heavy blow to the food supply of developing countries. This agricultural crisis on a global scale has drawn the intense attention of governments and the academic community worldwide to determine how to enhance the ability of agriculture to cope with meteorological disasters.

As a major agricultural country, agriculture holds a pivotal position in China’s national economy. With the rapid development of industrialization and urbanization, the risks of meteorological disasters faced by Chinese agriculture are escalating. From the droughts in northern regions to the flood disasters in southern regions, the scope of the impact of meteorological disasters on Chinese crops is constantly expanding, severely restricting the sustainable development of agriculture. To address this challenge, the Chinese government is actively promoting the modernization of agriculture and encouraging farmers to adopt more scientific and efficient pest and disease control measures. Among these, the cooperative pest and disease control behavior has emerged as an important development direction. As a new pest and disease management approach in China, the cooperative pest and disease control behavior is of great significance for reducing agricultural production risks. This is determined by China’s basic agricultural production situation: decentralized operation and management, small-scale operation areas, and a relatively severe differentiation among farmers [4,5]. The Chinese government allocates CNY 800 million annually as special funds for pest and disease control to support the unified prevention and control model. For example, in Yuzhou District, Guangxi, the characteristics of rice cultivation are that there are few large-scale farmers, while small and medium-sized farmers are in the majority. The farmland is scattered, and the degree of mechanization is low. Farmers entrust professional pest control service organizations to carry out drone-based cooperative pest control for 3000 mu of rice fields, effectively preventing and controlling the “three pests and two diseases” of rice. In Qiyang, Hunan, drone-based cooperative pest control has also been completed over an area of 520,000 mu. Traditional pest and disease control of crops relies on the efforts of individual farmers, which is technically demanding and costly. Against the backdrop of the implementation of the rural revitalization strategy and the promotion of “green agriculture”, the self-prevention and control model based on individual farmers is increasingly unable to meet the needs of environmentally friendly agriculture. The cooperative pest and disease control behavior refers to the choice behavior of farmers in a village entrusting the pest and disease control work to professional private agricultural social service organizations by paying fees. These organizations conduct unified pest and disease prediction, determine the timing of control and the types of pesticides, and implement unified control [6,7]. This approach can not only curb the spread of pests and diseases and reduce negative externalities but can also lower the control costs and enhance the overall effectiveness [8,9,10]. Despite the extensive publicity and guidance from relevant government departments highlighting the advantages of the cooperative pest and disease control behavior, the implementation of the unified prevention and control of crops remains slow, and the adoption rate among farmers is low [11].

However, currently, farmers still encounter numerous obstacles in the process of implementing the cooperative pest and disease control behavior [12,13]. On the one hand, the traditional concept of individual control is deeply ingrained, and farmers have a low level of awareness and acceptance of cooperative control. On the other hand, the cooperative control requires a high level of professional knowledge and technical skills, and farmers face difficulties in practical operation. In addition, issues such as information asymmetry and a shortage of funds also limit farmers’ enthusiasm for participating in cooperative control. These obstacles not only affect the effectiveness of pest and disease control but also increase the production costs of agriculture.

The existing literature on the factors influencing farmers’ cooperative pest and disease control behavior mainly covers farmers’ individual characteristics, family endowments, and external environmental characteristics. Individual characteristics can be mainly categorized into two aspects: one is demographic attributes, such as farmers’ age, gender, education level, and health status [14,15,16]; the other is cognitive characteristics. Existing studies have demonstrated that farmers’ cognitive level is a key factor influencing their behavioral choices. For example, higher social trust and a stronger perception of the effectiveness of unified prevention and control are associated with a higher enthusiasm and willingness among farmers to participate in cooperative pest and disease control [8,14]. Family endowments include off-farm employment, the size of cultivated land, and the shortage of labor within the family, all of which significantly impact farmers’ cooperative pest and disease control behavior. External environmental characteristics include market factors such as pesticide prices and agricultural product prices [17], as well as government policy support, both of which directly affect farmers’ production behaviors [6]. Overall, perspectives from individual characteristics, family endowment characteristics, agricultural product prices, social division of labor, social trust, risk preferences, and attitudes have been used to explain the slow implementation and low adoption rate of the unified prevention and control of crops. However, the most prominent reason is attributed to insufficient demand.

Meteorological disaster shocks, as an important factor affecting agricultural production, have a profound impact on farmers’ cooperative pest and disease control behavior. Frequent meteorological disasters will change the occurrence patterns and spread ranges of pests and diseases, rendering the original control strategies of farmers ineffective. Under drought conditions, the reproduction rate of some pests and diseases accelerates, causing more severe harm to crops, which prompts farmers to seek more effective control methods and, thus, increases the demand for cooperative pest and disease control. Risk perception, as an important basis for farmers’ decision making, also significantly influences their cooperative pest and disease control behavior [18,19]. The higher the degree of farmers’ perception of meteorological disaster risks, the more inclined they are to take proactive control measures to reduce losses. When farmers are aware that meteorological disasters may lead to large-scale outbreaks of pests and diseases, it intensifies their perception of the existing risks and the perception of potential future risk losses. In order to reduce production risks and future losses, they are more likely to choose to cooperate with other farmers to jointly carry out pest and disease control. This behavior decision making based on risk perception is of great significance for optimizing the allocation of agricultural production resources and improving agricultural production efficiency. Moreover, by promoting farmers to adopt cooperative pest and disease control behavior, it can effectively increase crop yields, ensure food security, and contribute to the achievement of the “Zero Hunger” goal. At the same time, the cooperative control behavior can also reduce the use of pesticides, lower agricultural non-point source pollution, and help achieve sustainable development goals such as “Clean Water and Sanitation” and “Climate Action” [9,16]. Therefore, in-depth research on the impact of meteorological disaster shocks and risk perception on farmers’ cooperative pest and disease control behavior has important practical significance for promoting the sustainable development of global agriculture.

This study not only focuses on the impact of meteorological disaster shocks and risk perception on Chinese farmers’ cooperative pest and disease control behavior but also explores its applicability in other regions. As a major agricultural country, China has rich agricultural production experience and diverse agricultural ecosystems, and the research results have certain reference value for the agricultural development of other developing countries and even the entire world. In regions with similar agricultural ecosystems, they can draw on the experience of Chinese farmers in dealing with meteorological disasters and pest and disease control to optimize agricultural production strategies.

In conclusion, this study uses the survey data of 613 farmers from two provinces in China to investigate the impact of meteorological disaster shocks and risk perception on farmers’ cooperative pest and disease control behavior and to analyze the mediating role of risk perception in the impact of meteorological disaster shocks on farmers’ cooperative pest and disease control behavior, providing a new perspective on farmers’ decision-making mechanisms in the face of meteorological disasters. In addition, this paper also studies the heterogeneity of farmers’ cooperative pest and disease control behavior and examines the behavioral differences among farmers of different scales in coping with meteorological disasters.

2. Theoretical Analysis and Hypotheses

2.1. Impact of Meteorological Disaster Shocks on Farmers’ Cooperative Pest and Disease Control Behavior

Drawing on behavioral economics and cost–benefit theories, scholars have identified that external environment and information, basic characteristics of farmers, psychological cognition, and operational benefits are key factors influencing farmers’ decision-making behaviors [20,21]. Generally, aiming to maximize profits, farmers first consider the positive impacts of external factors. For example, the use of digital technology to reduce information costs or enhance technical knowledge levels can positively influence farmers’ decision making [22,23]. However, meteorological disaster shocks bring negative impacts, such as reduced agricultural yields and threats to family sustainable operations. Farmers carefully weigh risks and benefits, which can stimulate their cooperative pest and disease control behavior.

After experiencing droughts and floods, pest and disease outbreaks intensify, threatening crop yields and family incomes [24]. This changes farmers’ expected income, cost perception, and risk assessment levels. Therefore, after suffering from meteorological disaster shocks, farmers may actively engage in cooperative pest and disease control behavior to mitigate future risks. The mechanisms include the following.

Sharing the cost of risk-bearing: Based on prospect theory [25], under the impact of meteorological disasters, the economic situation of farmers is often affected, and they may face great pressure when bearing the cost of pest and disease control alone. Cooperative control, however, can achieve cost sharing, which is very attractive to farmers. When multiple farmers cooperate to purchase pesticides, lease large-scale control equipment, etc., they can enjoy the price discounts brought by bulk purchases, reducing the unit control cost. At the same time, cooperation can also enable the sharing of labor resources, alleviating the labor shortage problem that may be caused by meteorological disasters, and further reducing the control cost. For example, after a flood disaster, some farmers may be short of labor for pest and disease control as they are busy rescuing crops and restoring production. Through cooperation, the labor force of other farmers can be allocated to jointly complete the control work, and the cost of hiring additional labor can be shared.

Reducing technical uncertainty and saving information costs: Under meteorological disaster shocks, farmers proactively strengthen connections and cooperation with other farmers in the same village, increasing trust and overcoming panic. They seek mutual assistance in agricultural production materials and psychological support, enhancing information exchange in cooperative pest and disease control. Higher information transparency helps farmers obtain more comprehensive information, reducing technical uncertainty and saving information costs [26,27].

Alleviating family labor constraints: (1) Labor resource integration. Through cooperative pest and disease control, farmers can pool the labor forces of individual households and make unified deployment and arrangements [14]. For example, during the critical period of pest and disease control, the limited labor force can be allocated to the farmland areas that are most severely affected by disasters or where pests and diseases are most likely to break out, improving the utilization efficiency of labor and avoiding the situation where pest and disease control is delayed due to the insufficient labor force of a single household. (2) Professional division of labor improves efficiency. In cooperative pest and disease control, farmers can carry out professional division of labor according to their respective skills and experiences. Some farmers are good at operating pest control equipment, and some farmers have more experience in identifying and judging pests and diseases. Through reasonable division of labor, the overall efficiency of pest and disease control can be improved, and the demand for the number of labor forces can be relatively reduced. Taking the use of drones for pest and disease control as an example, having farmers who are professional in operating drones carry out unified operations is more labor-saving than each household using traditional methods for control. (3) Sharing labor costs. Cooperative pest and disease control can reduce the cost of individual farmers hiring additional labor forces by sharing labor resources. When multiple farmers cooperate, the fees for jointly hiring professional pest and disease control teams or temporary workers can be shared among them, relieving the economic burden of each household. At the same time, during the cooperation process, farmers can also help each other without having to pay additional rewards, further saving labor costs.

2.2. Risk Perception and Farmers’ Cooperative Pest and Disease Control Behavior Under Meteorological Disaster Shocks

After meteorological disaster events such as droughts, heavy rains, and floods, crops may experience lodging, drying, freezing damage, and increased pest and disease outbreaks. Indirectly, this can also degrade arable land quality. These hazards significantly impact individual farmers’ cognition [28,29], forming a “stimulus source”. This prompts farmers to make psychological judgments about the scope [30], manner, and nature of the hazards based on their own cognition and experience, forming the “organism” variable in the SOR response model, that is, generating farmers’ risk perception of meteorological disasters [31,32,33].

The mediating role of risk existence perception: (1) Adjusting production decisions. After farmers experience the shock of meteorological disasters, their perception of the existence of risks will be enhanced [34]. They will realize that agricultural production faces numerous uncertainties, which will prompt farmers to re-evaluate the likelihood of the occurrence of pests and diseases. As a result, they will change their production decisions, shifting from originally not paying much attention to pest and disease control to actively considering various prevention and control measures, including cooperative pest and disease control. (2) Strengthening the willingness to cooperate. The perception of risk existence will make farmers aware of the limitations of their own ability to cope with risks, and then they will seek more effective ways to deal with risks. They will find that collaborating with other farmers in pest and disease control can integrate resources, share information and technologies, and improve their ability to deal with pests and diseases. For example, in arid regions, after farmers perceive the risk that droughts may trigger insect disasters, they are more willing to cooperate with their neighbors and jointly hire professional pest and disease control teams to enhance their ability to resist potential pest and disease risks. (3) Promoting resource integration. In cooperative pest and disease control, farmers can pool their respective resources such as funds, equipment, and labor forces. They can purchase prevention and control equipment, pesticides, etc., uniformly, which improves the efficiency of resource utilization and realizes economies of scale, enabling them to better deal with possible pests and diseases. This is also an important reason why farmers choose cooperative pest and disease control behavior under the perception of risk existence. In this way, an influence path of “stimulated by meteorological disaster shocks → generating the perception of risk existence → farmers’ cooperative pest and disease control behavior” is formed [35].

The mediating role of risk loss perception: (1) Increasing prevention and control investment. Farmers who have experienced disasters such as droughts are well aware of the serious losses that pests and diseases may cause after the disaster. They will not hesitate to increase costs to purchase better pesticides, invite professionals to provide guidance on prevention and control, and even take the initiative to cooperate with other farmers to jointly bear the prevention and control costs in order to reduce risk losses. (2) Optimizing prevention and control technologies. The perception of risk losses will prompt farmers to seek more effective pest and disease prevention and control technologies and methods, and cooperative prevention and control often provides such a platform [36]. Multiple farmers collaborating can invite experts to conduct technical training and jointly introduce advanced prevention and control equipment and technologies, such as biological control technologies and intelligent monitoring devices. (3) Strengthening risk prevention. A strong perception of risk losses will make farmers pay more attention to long-term risk prevention, rather than just short-term prevention and control effects. By participating in cooperative pest and disease control, farmers can establish a long-term pest and disease monitoring and prevention and control system, and jointly formulate prevention and control plans, so as to reduce the risk losses caused by pests and diseases triggered by meteorological disasters in the future and ensure the long-term stability of agricultural production. In this way, an influence path of “stimulated by meteorological disaster shocks → generating the perception of risk losses → farmers’ cooperative pest and disease control behavior” is formed.

3. Materials and Methods

3.1. Data Source

The data used in this paper were collected during field surveys conducted by the research team in Gansu and Ningxia from July to December 2021. The reasons for selecting maize growers in Ningxia and Gansu are as follows: First, in the regions of Gansu and Ningxia, maize is one of the main food crops. Compared with other crops, maize has better drought tolerance and adaptability, and it has a wide planting area. It is a representative crop variety in the local area, occupying an important position in the local agricultural production structure and having a significant impact on the local agricultural economic development and farmers’ income. Studying issues related to maize cultivation is of great significance for ensuring regional food security and promoting agricultural economic development. Second, Gansu and Ningxia are located in the inland areas of the northwest of China, with complex climatic conditions. Meteorological disasters such as droughts, heavy rains, and sandstorms occur frequently. These meteorological disasters have a significant impact on maize cultivation, easily causing problems such as hindered growth and development of maize and reduced yields, providing a typical research scenario for studying the impact of meteorological disaster shocks on crops. In addition, there is a wide variety of maize pests and diseases. For example, the Asian corn borer, northern corn leaf blight, etc., are quite common in maize cultivation in Gansu and Ningxia regions, and there is a certain correlation between the occurrence of these pests and diseases and meteorological disasters, which is convenient for studying the relationship among meteorological disaster shocks, risk perception, and farmers’ cooperative pest and disease control behaviors. A combination of typical sampling and random sampling methods was employed. The specific steps are as follows: First, based on factors such as the occurrence of meteorological disasters and agricultural economic development in the northwest region, three counties (cities) were selected: Linze County in Gansu Province, Minle County in Gansu Province, and Qingtongxia City (a county-level city) in Ningxia Hui Autonomous Region. Second, nine townships were randomly selected, and in each township, 2–4 sample villages were randomly chosen. Finally, 20–30 maize growers were randomly surveyed in each sample village. A total of 613 valid questionnaires were obtained, with 273 households from Gansu Province and 340 households from Ningxia Hui Autonomous Region. This paper utilizes STATA17.0 (StataCorp, College Station, TX, USA, www.stata.com) software for regression.

3.2. Variable Selection and Description

Cooperative pest and disease control behavior: Cooperative pest and disease control behavior refers to the cooperative actions among farmers or involving multiple parties in agricultural production to jointly control the losses caused by pests and diseases to crop yields and quality. In this study, the degree of cooperative pest and disease control behavior was measured by whether farmers adopted full or semi-package cooperative pest and disease control services, with a value of “1” indicating adoption and “0” indicating non-adoption. According to

Table 1. Assignment and descriptive statistics of sub-indicators for risk perception.

Composite Indicator	Definition Explanation	Mean	Standard Deviation
Risk Existence Perception	Future three-year intensification of drought.	3.4241	1.1643
	Future three-year intensification of heavy rain and flood.	2.6394	1.1794
	Future three-year intensification of low-temperature freezing damage.	3.2936	1.1297
	Future three-year intensification of strong wind and hail.	3.3572	0.1381
Risk Loss Perception	Meteorological disasters are not conducive to agricultural production investment.	4.1664	0.7702
	Meteorological disasters reduce crop yields.	4.2495	0.7721
	Meteorological disasters reduce crop quality.	4.0897	0.8498
	Meteorological disasters reduce agricultural income.	4.2724	0.7791

, the proportion of farmers engaging in cooperative pest and disease control behavior accounted for 13.21% of the total number of farmers.

Meteorological disaster shocks: When studying the factor of meteorological disaster shocks, the research by Xu et al. (2023) and Wei et al. (2023) was referenced [37,38]. Considering the lag characteristic of farmers’ choice of cooperative pest and disease control behavior, two variables were used to represent meteorological disaster shocks: whether farmers had suffered from drought and flood disasters in the past three years and the number of times they had suffered. Specifically, farmers were asked “Have you suffered from drought and flood disasters in the past three years?” If the answer was “yes”, the value was assigned as “1”; otherwise, it was “0”. According to the statistics in

Table 2. Assignment and descriptive statistics of main variables.

Variable Name	Definition and Assignment Explanation	Mean	Standard Deviation
Dependent Variable
Cooperative Pest and Disease Control Behavior	Whether farmers choose cooperative pest and disease control behavior: 1 = yes, 0 = no.	0.1321	0.3389
Core Explanatory Variable
Meteorological Disaster Shocks	Whether suffered from meteorological disaster shocks in the past three years.	0.1647	0.3712
Frequency of Meteorological Disaster Shocks	Number of times suffered from meteorological disaster shocks in the past three years.	0.2773	0.6917
Mediating Variable
Risk Existence Perception	Perception of meteorological disaster risk existence: calculated by entropy method.	3.1578	0.7771
Risk Loss Perception	Perception of meteorological disaster risk loss: calculated by entropy method.	4.1380	0.6822
Instrumental Variable
Probability of Other Villagers in the Same Township Suffering from Meteorological Disaster Shocks	Probability of other villagers in the same township suffering from meteorological disaster shocks.	0.2171	0.2562
Control Variables
Decision-Maker’s Age	Age of the decision-maker (years).	56.8303	9.6989
Decision-Maker’s Education Level	Education level of the decision-maker: illiterate = 1, primary school = 2, junior high school and secondary vocational school = 3, high school = 4, college and above = 5.	2.6508	0.7533
Risk Preference	If you have a sum of money, which project would you like to invest in: high risk high return = 1, moderate risk moderate return = 2, low risk low return = 3, risk-free stable return = 4.	3.3393	0.8813
Number of Family Laborers	Number of family members with labor capacity (persons).	3.1027	1.0637
Proportion of Family Agricultural Income	Proportion of family agricultural income in total income (%).	0.6438	0.2963
Maize Planting Area	Family maize planting area (mu).	15.9832	15.1101
Maize Sales Price	Maize sales price of the year (CNY).	1.3412	0.2326
Type of Cultivated Land	Type of family maize planting cultivated land: flat land = 1, mountainous area = 0.	0.9787	0.1441
Government Support	Whether the government provides early warning information services, various funds, technologies, and production materials for meteorological disasters: 1 = yes, 0 = no.	0.3898	0.4881
Participation in Industrial Organizations	Whether farmers participate in industrial organizations: 1 = yes, 0 = no.	0.4176	0.4935
Social Network	Number of contacts in the investigator’s mobile phone (persons).	155.5905	319.5325

, 16.47% of the farmers have suffered from drought or flood disasters in the past three years.

Risk perception: Risk perception refers to farmers’ level of awareness regarding the possibility, severity, and potential losses associated with meteorological disaster risks. It is a psychological response activity. Therefore, this paper categorizes risk perception into risk existence perception and risk loss perception [34]. Based on the method of measuring indicators by Zhao and Yu (2019), a composite index was used to reflect risk existence perception and risk loss perception caused by meteorological disaster risks. Specifically, four sub-indicators were measured using the entropy method [39]. These indicators were assigned values using a Likert five-point scale, where 1 indicates strong disagreement and 5 indicates strong agreement. The specific indicators are shown in in Table 1.

Instrumental variable: Referring to the instrumental variable selection methods by Sheng et al. (2021) and Li et al. (2022), the probability of other villagers in the same township suffering from meteorological disaster shocks was selected as the instrumental variable [40,41].

Control variables: Individual characteristics include the age, education level, and risk preference of the household decision maker. Family endowment characteristics include the number of family laborers, the proportion of family agricultural income, the type of cultivated land, the area of crop cultivation, social networks, participation in industrial organizations, etc. External environmental characteristics include crop sales prices and government support. For details, see Table 2.

3.3. Model Specification

3.3.1. Benchmark Regression Model

When the dependent variable is “cooperative pest and disease control behavior”, it is a binary choice problem, and the Probit model is used for estimation. The benchmark model is set as follows:

$${\mathrm{M}\mathrm{P}}_{\mathrm{i}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{\mathrm{M}\mathrm{D}}_{\mathrm{i}}+\mathsf{\beta }{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}_{\mathrm{i}}+{\mathsf{\epsilon }}_{\mathrm{i}}$$

(1)

$${\mathrm{M}\mathrm{P}}_{\mathrm{i}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_2{\mathrm{N}\mathrm{M}\mathrm{D}}_{\mathrm{i}}+\mathsf{\beta }{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}_{\mathrm{i}}+{\mathsf{\epsilon }}_{\mathrm{i}}$$

(2)

$${\mathrm{M}\mathrm{P}}_{\mathrm{i}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{\mathrm{N}\mathrm{M}\mathrm{D}}_{\mathrm{i}}+{\mathsf{\alpha }}_2{{\mathrm{N}\mathrm{M}\mathrm{D}}^2}_{\mathrm{i}}+\mathsf{\beta }{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}_{\mathrm{i}}+{\mathsf{\epsilon }}_{\mathrm{i}}$$

(3)

where ${\mathrm{M}\mathrm{P}}_{\mathrm{i}}$ represents the cooperative behavior of farmers in pest and disease control; ${\mathrm{M}\mathrm{D}}_{\mathrm{i}}$, ${\mathrm{N}\mathrm{M}\mathrm{D}}_{\mathrm{i}}$ and ${{\mathrm{N}\mathrm{M}\mathrm{D}}^2}_{\mathrm{i}}$ represent the impact of meteorological disasters, the number of meteorological disaster shocks, and the squared term of the number of shocks, respectively; control_i represents other control variables that may affect the cooperative behavior of farmers in pest and disease control; ${\mathsf{\epsilon }}_{\mathrm{i}}$ is the random disturbance term. ${\mathsf{\alpha }}_0$, ${\mathsf{\alpha }}_1$, ${\mathsf{\alpha }}_2$, and $\mathsf{\beta }$ represent the regression coefficients corresponding to the explanatory variables.

3.3.2. Mediating Effect Model (Bootstrap Test)

To further verify whether meteorological disaster shocks can affect farmers’ cooperative pest and disease control behavior through influencing risk perception, the following mediating effect analysis model is set up:

$${\mathrm{L}\mathrm{O}}_{\mathrm{i}}={\mathsf{\rho }}_0+{\mathsf{\rho }}_1{\mathrm{M}\mathrm{D}}_{\mathrm{i}}+\mathsf{\beta }{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}_{\mathrm{i}}+{\mathsf{\epsilon }}_{\mathrm{i}}$$

(4)

$${\mathrm{M}\mathrm{P}}_{\mathrm{i}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{\mathrm{M}\mathrm{D}}_{\mathrm{i}}+\mathrm{h}{\mathrm{L}\mathrm{O}}_{\mathrm{i}}+\mathsf{\beta }{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}_{\mathrm{i}}+{\mathsf{\epsilon }}_{\mathrm{i}}$$

(5)

where ${\mathrm{L}\mathrm{O}}_{\mathrm{i}}$ represents the level of farmers’ risk perception, and h is the regression coefficient of the risk perception variable. The meanings of other explanatory variables are the same as above. Referring to the mediating effect test process by Wen (2014), the existence of the mediating effect of risk perception is tested, and the significance of the mediating effect is tested using the stepwise regression method and the KHB method [42].

4. Results

4.1. Bayesian Information Criterion Statistical Test

Before estimating the model, we conducted a Bayesian information criterion statistical test on all the selected variables [43]. This test is used to compare and select among different statistical models. Its purpose is to penalize the complexity of the model while taking into account the goodness-of-fit of the model, so as to avoid overfitting. It can be seen from

Table 3. Bayesian Information Criterion statistical test.

Name	Parameter Name	Parameter Value
Model Parameter Settings	Data Preprocessing	None
	Proportion of Training Set	0.8
	Smoothing Processing (alpha Value)	1.0
	Type of Feature Distribution	Gaussian Distribution
Model Evaluation Effect	Accuracy	85.366%
	Precision (Comprehensive)	86.329%
	Recall (Comprehensive)	85.366%
	F1-score	0.858

that the accuracy of the final model on the test set is 85.37%, the precision (comprehensive) is 86.33%, the recall (comprehensive) is 85.37%, and the F1-score (comprehensive) is 0.86. The performance of the model is acceptable.

The ROC curve analysis was carried out, and the AUC index values of the training set and the test set were obtained as 0.801 and 0.784, respectively (As shown in Figure 1). The AUC values of the training set and the test set are quite close, which indicates that there is no obvious overfitting phenomenon in the model, and the model has good stability and generalization ability.

4.2. Estimation Results of Benchmark Regression Model

The estimation results using Stata 17.0 software are shown in

Table 4. Estimation results of benchmark regression model.

Variable Name	Cooperative Pest and Disease Control Behavior
	Model (1)	Model (2)	Model (3)
Meteorological Disaster Shocks	0.6786 *** (0.1695)
Frequency of Meteorological Disaster Shocks		0.3712 *** (0.0877)	0.6458 ** (0.2802)
Frequency Squared			−0.1105 (0.1074)
Decision Maker’s Age	0.0092 (0.0087)	0.0095 (0.0087)	0.0094 (0.0088)
Decision Maker’s Education Level	0.0243 (0.0971)	0.0124 (0.0974)	0.0245 (0.0981)
Risk Preference	−0.1340 * (0.0802)	−0.1087 (0.0815)	−0.1105 (0.0815)
Number of Family Laborers	−0.0319 (0.0658)	−0.0396 (0.0657)	−0.0373 (0.0660)
Proportion of Family Agricultural Income	0.0481 (0.2766)	−0.0295 (0.2759)	−0.0040 (0.2783)
Maize Planting Area	0.0063 (0.0048)	0.0077 (0.0047)	0.0071 (0.0048)
Maize Sales Price	0.6453 * (0.3339)	0.6463 * (0.3355)	0.6245 * (0.3364)
Type of Cultivated Land	−1.1583 *** (0.3818)	−1.2061 *** (0.3883)	−1.2094 *** (0.3847)
Government Support	0.0222 (0.1601)	0.0188 (0.1602)	0.0020 (0.1615)
Participation in Industrial Organizations	0.4238 ** (0.1708)	0.4062 ** (0.1711)	0.4088 ** (0.1715)
Social Network	0.0009 *** (0.0003)	0.0009 *** (0.0003)	0.0010 *** (0.0003)
Constant Term	−1.6651 * (0.9362)	−1.6208 * (0.9448)	−1.6416 * (0.9437)
LR chi2(15)	111.34 ***	113.01 ***	114.06 ***
Pseudo R2	0.2326	0.2361	0.2383

Note: *, **, and *** indicate significance at the 10%, 5%, and 1% levels, respectively. The numbers in parentheses are standard errors.

, and were used to empirically estimate the impact of meteorological disaster shocks on farmers’ cooperative pest and disease control behavior. In Model (1), only the impact of meteorological disaster shocks and control variables on farmers’ cooperative pest and disease control behavior was considered. In Model (2), only the impact of the frequency of meteorological disaster shocks on farmers’ cooperative pest and disease control behavior was considered. In Model (3), both the frequency of meteorological disaster shocks and the squared frequency were considered. The empirical results show that meteorological disaster shocks have a significant positive impact on farmers’ cooperative pest and disease control behavior at the 1% statistical level. This indicates that as the frequency of meteorological disaster shocks increases, farmers are more likely to choose cooperative pest and disease control behavior.

Among the control variables, the type of cultivated land has a significant negative relationship with farmers’ cooperative pest and disease control behavior at the 1% statistical level. This suggests that on mountainous cultivated land, farmers are more likely to choose cooperative pest and disease control behavior. Participation in industrial organizations and social networks have a significant positive impact on farmers’ cooperative pest and disease control behavior at the 5% and 1% statistical levels, respectively. This indicates that farmers who participate in industrial organizations and have wider social networks are more likely to adopt cooperative pest and disease control behavior measures. This may be because the larger the social network of farmers, the more frequent the information exchange, the richer the agricultural knowledge accumulated, and the higher the level of agricultural production management, which reduces information asymmetry. The maize sales price has a significant positive impact on farmers’ cooperative pest and disease control behavior at the 10% statistical level. This suggests that the higher the maize sales price, the more likely farmers are to choose cooperative pest and disease control behavior. This is because the higher the maize sales price, the better the market prospects for maize, and the more encouraged farmers are to choose cooperative pest and disease control behavior.

4.3. Estimation Results of Addressing Endogeneity

To overcome potential endogeneity issues, following the research methods of Sheng et al. (2021), the “probability of other villagers in the same township suffering from meteorological disaster shocks” was used as an instrumental variable for IVprobit model regression [40]. Referring to the research by Yuan (2018), identification tests and weak instrumental variable tests were conducted from multiple angles to rule out the possibility of weak instrumental variables [44]. In the estimation results of addressing endogeneity of meteorological disaster shocks on cooperative pest and disease control behavior (see

Table 5. Estimation results of addressing endogeneity.

Variable Name	Meteorological Disaster Shocks	Frequency of Meteorological Disaster Shocks	Frequency Squared	Cooperative Pest and Disease Control Behavior
	First Stage	First Stage	First Stage	Second Stage	Second Stage	Second Stage
Meteorological Disaster Shocks				1.6290 *** (0.2653)
Frequency of Meteorological Disaster Shocks					1.0989 *** (0.1501)	2.5757 *** (0.4753)
Frequency Squared			0.3490 *** (0.0046)			−0.8220 *** (0.1794)
Temperature Difference	0.9006 *** (0.0551)	1.2283 *** (0.1127)	0.4896 *** (0.0364)
Decision Maker’s Age	0.0031 ** (0.0014)	0.0054 * (0.0028)	0.0010 (0.0009)	0.0063 (0.0085)	0.0046 (0.0081)	0.0075 (0.0084)
Decision Maker’s Education Level	−0.0151 (0.0175)	−0.0089 (0.0357)	−0.0271 ** (0.0111)	0.0317 (0.0933)	0.0123 (0.0892)	0.0674 (0.0931)
Risk Preference	−0.0429 *** (0.0145)	−0.1279 *** (0.0297)	−0.0172 * (0.0094)	−0.0991 (0.0778)	−0.0087 (0.0780)	−0.0887 (0.0774)
Number of Family Laborers	−0.0008 (0.0113)	0.0178 (0.0232)	−0.0049 (0.0072)	−0.0485 (0.0633)	−0.0661 (0.0603)	−0.0420 (0.0622)
Proportion of Agricultural Income	−0.0026 (0.0424)	0.0635 (0.0868)	−0.0189 (0.0270)	0.2269 (0.2752)	0.1034 (0.2583)	0.2125 (0.2723)
Maize Planting Area	−0.0006 (0.0010)	−0.0008 (0.0020)	−0.0008 (0.0006)	−0.0003 (0.0050)	−0.0002 (0.0048)	0.0009 (0.0049)
Maize Sales Price	−0.0469 (0.0571)	−0.0241 (0.1168)	0.0180 (0.0363)	0.5967 * (0.3214)	0.5157 * (0.3096)	0.4819 (0.3202)
Type of Cultivated Land	0.0421 (0.0829)	0.1193 (0.1696)	0.0568 (0.0528)	−1.0931 *** (0.3688)	−1.0882 *** (0.3627)	−1.1725 *** (0.3674)
Government Support	0.0313 (0.0273)	0.0811 (0.0559)	0.0289 * (0.0174)	−0.0973 (0.1581)	−0.1373 (0.1509)	−0.1363 (0.1565)
Participation in Industrial Organizations	−0.0183 (0.0282)	0.0102 (0.0577)	−0.0049 (0.0180)	0.5236 *** (0.1690)	0.4445 *** (0.1594)	0.4923 *** (0.1654)
Social Network	−0.0000 (0.0000)	−0.0001 * (0.0001)	−0.0000 (0.0000)	0.0009 *** (0.0003)	0.0009 *** (0.0002)	0.0009 *** (0.0003)
Constant Term	0.0131 (0.1623)	−0.0299 (0.3321)	0.0019 (0.1033)	−1.7208 * (0.9074)	−1.5210 * (0.8707)	−1.6372 * (0.8985)
Endogeneity Wald Test Value				14.97 ***	17.65 ***	16.15 ***
Instrumental Variable Wald Test Value				27.45 ***	27.18 ***	20.40 ***
Instrumental Variable Wald Test Value				25.69 ***	23.64 ***	18.47 ***

Note: *, **, and *** indicate significance at the 10%, 5%, and 1% levels, respectively. The numbers in parentheses are standard errors.

), the Wald endogeneity test values were 14.97, 17.65, and 16.15, all of which were significant at the 1% statistical level. This rejects the null hypothesis that meteorological disaster shocks and their frequency are exogenous variables, indicating the presence of endogeneity and the appropriateness of using the instrumental variable method. In Table 5, the coefficients in the first stage were 0.9009, 0.12283, and 0.4896, all of which were significant at the 1% statistical level. The AR test values were 27.45, 27.18, and 20.40, all of which were significant at the 1% statistical level; the Wald test values were 25.69, 23.64, and 18.47, all of which were significant at the 1% statistical level. This indicates that the identification test was passed and there was no weak instrumental variable problem. In summary, using the “probability of other villagers in the same township suffering from meteorological disaster shocks” as an instrumental variable and applying the IVprobit model regression effectively addressed the endogeneity issue in this chapter.

4.4. Mechanism Test

Using the mediating effect model, this study tested the impact mechanism of meteorological disaster shocks on farmers’ cooperative pest and disease control behavior.

Table 6. Mediating effect test results of risk existence perception: Bootstrap test method.

Variable Name	Model (1)	Model (2)	Variable Name	Model (3)	Model (4)
	Risk Existence Perception	Cooperative Pest and Disease Control Behavior		Risk Existence Perception	Cooperative Pest and Disease Control Behavior
Meteorological Disaster Shocks	0.3671 *** (0.0838)	0.6317 *** (0.1748)	Frequency of Meteorological Disaster Shocks	0.1442 *** (0.0458)	0.3519 *** (0.0894)
Risk existence Perception		0.1144 (0.1053)	Risk existence Perception		0.1328 (0.1036)
Mediating Effect: Meteorological Disaster Shocks		0.0071 (0.0052)	Mediating Effect: Frequency of Meteorological Disaster Shocks		0.0029 (0.0025)
Confidence Interval		[−0.0003, 0.0207]	Confidence Interval		[−0.0009, 0.0083]
R2	0.1097	0.2486	R2	0.1001	0.2351

Note: *** indicate significance at the 1% levels, respectively. The numbers in parentheses are standard errors.

,

Table 7. Results of the mediating effect test of the perceived risk: KHB method.

Variable Name	Risk Existence Perception	Cooperative Pest and Disease Control Behavior	Variable Name	Risk Existence Perception	Cooperative Pest and Disease Control Behavior
Meteorological Disaster Shocks	Total Effect	0.6735 *** (0.1699)	Frequency of Meteorological Disaster Shocks	Total Effect	0.3705 *** (0.0883)
	Direct Effect	0.6317 *** (0.1748)		Direct Effect	0.3519 *** (0.0894)
	Indirect Effect	0.0417 (0.0396)		Indirect Effect	0.0186 (0.0157)
	Mediating Effect	Significant		Mediating Effect	Not Significant

Note: *** indicate significance at the1% levels, respectively. The numbers in parentheses are standard errors.

,

Table 8. Mediating effect test results of risk loss perception: Bootstrap test method.

Variable Name	Model (1)	Model (2)	Variable Name	Model (3)	Model (4)
	Risk Loss Perception	Cooperative Pest and Disease Control Behavior		Risk Loss Perception	Cooperative Pest and Disease Control Behavior
Meteorological Disaster Shocks	0.0890 *** (0.0210)	0.1333 *** (0.0352)	Frequency of Meteorological Disaster Shocks	0.1511 *** (0.0394)	0.0847 *** (0.0187)
Risk Loss Perception		0.0533 ** (0.0184)	Risk Loss Perception		0.0524 *** (0.0183)
Mediating Effect: Meteorological Disaster Shocks		0.0173 * (0.0061)	Mediating Effect: Frequency of Meteorological Disaster Shocks		0.0083 ** (0.0030)
Confidence Interval		[0.0050, 0.0300]	Confidence Interval		[0.0034, 0.0131]
R2	0.1396	0.2314	R2	0.1396	0.2314

Note: *, **, and *** indicate significance at the 10%, 5%, and 1% levels, respectively. The numbers in parentheses are standard errors.

and

Table 9. Results of the mediating effect test of the perceived risk loss:: KHB method.

Variable Name	Risk Loss Perception	Cooperative Pest and Disease Control Behavior	Variable Name	Risk Loss Perception	Cooperative Pest and Disease Control Behavior
Meteorological Disaster Shocks	Total Effect	0.4216 *** (0.1306)	Meteorological Disaster Shocks	Total Effect	0.4216 *** (0.1306)
	Direct Effect	0.3729 *** (0.1322)		Direct Effect	0.3729 *** (0.1322)
	Indirect Effect	0.0486 *** (0.0186)		Indirect Effect	0.0486 *** (0.0186)
	Mediating Effect	Significant		Mediating Effect	Significant
	Mediating Effect Proportion	11.52%		Mediating Effect Proportion	11.52%

Note: *** indicate significance at the 1% levels, respectively. The numbers in parentheses are standard errors.

show the results of the mechanism test using the stepwise regression method and the KHB method [45].

Based on the estimation results in Table 6 and Table 7, both weather disaster impacts and the frequency of weather disaster impacts in Models (1) and (3) have a positive and significant impact on the perception of risk existence at the 1% statistical level, indicating that they enhance the perception of risk existence. The regression results of Models (2) and (4) show that the perception of risk existence has no significant impact on farmers’ cooperative pest control behaviors. This suggests that the perception of risk existence does not mediate the influence of weather disaster impacts and their frequency on farmers’ cooperative pest control behaviors.

According to the estimation results in Table 8 and Table 9, both weather disaster impacts and the frequency of weather disaster impacts in Models (1) and (3) have a positive and significant impact on the perception of risk loss at the 1% statistical level, indicating that they enhance the perception of risk loss. The regression results of Models (2) and (4) show that the impact of the perception of risk loss on farmers’ cooperative pest control behaviors is significantly positive at the 5% and 1% statistical levels, respectively. This suggests that the perception of risk loss mediates the influence of weather disaster impacts and their frequency on farmers’ cooperative pest control behaviors.

The dependent variables in this chapter are all binary variables. To avoid bias in the estimation results, the KHB method, which is suitable for discrete dependent variables, is once again selected for mediation effect decomposition estimation. As shown in Table 9, in the process of weather disaster impacts influencing farmers’ cooperative pest control behaviors, the indirect effects of the perception of risk loss pass the significance test, and the coefficients are all positive, indicating that weather disaster impacts can indirectly promote farmers’ cooperative pest control behaviors through the perception of risk loss. The decomposition estimation results of the KHB method are consistent with those of the stepwise regression method. According to Table 9, the indirect effects of the perception of risk loss account for 11.52% and 11.97% of the total effects of weather disaster impacts and their frequency on farmers’ cooperative pest control behaviors, respectively

Combining the SOR model in decision-making situation theory with the mediation effect regression results for mechanism analysis: Meteorological factors are an important part of the natural resource conditions on which the development of the agricultural industry depends. Weather disaster impacts affect crop growth and agricultural output, reducing agricultural income and making agricultural producers more sensitive to meteorological factors, leading to the cognition or perception of risk loss from weather disasters. Meanwhile, agricultural business entities are boundedly rational. In the pursuit of risk minimization and expected utility maximization, they make value judgments about cooperative pest control behaviors and choose the best preventive measures based on their production and operation conditions and resource endowments to reduce the probability of weather disaster risks and losses after they occur, maintain or increase agricultural income, and reduce agricultural production risks.

5. Heterogeneity Analysis and Robustness Test

5.1. Heterogeneity Analysis

To fully understand the differences in farmers’ cooperative pest and disease control behavior under meteorological disaster shocks among different types of farmers, the Probit model was used for estimation. Referring to the classification methods by Niu et al. (2022) and Wang et al. (2019), the decision maker’s education level and planting scale were grouped into primary school and below, junior high school and above, small-scale (10 mu and below), and large-scale (more than 10 mu) groups [46,47]. These two indicators can, to some extent, represent the family’s human capital and dependence on the agricultural industry, testing the impact of meteorological disaster shocks on the cooperative pest and disease control behavior of different types of farmers.

As shown in

Table 10. Regression results of different education groups.

Variable Name	Cooperative Pest and Disease Control Behavior
	Model (1) Primary School and Below	Model (2) Junior High School and Above
Meteorological Disaster Shocks	0.2590 (0.2708)	1.0342 *** (0.2477)
Control Variables	Controlled	Controlled
LR chi2 (15)	42.74 ***	86.29 ***
Pseudo R2	0.1968	0.3368
Variable Name	Cooperative Pest and Disease Control Behavior
	Model (3) Primary School and Below	Model (4) Junior High School and Above
Frequency of Meteorological Disaster Shocks	0.2070 (0.1467)	0.5087 *** (0.1213)
Control Variables	Controlled	Controlled
LR chi2 (15)	43.76 ***	86.29 ***
Pseudo R2	0.2015	0.3368
Variable Name	Cooperative Pest and Disease Control Behavior
	Model (5) Primary School and Below	Model (6) Junior High School and Above
Frequency of Meteorological Disaster Shocks	0.1232 (0.4440)	0.9902 ** (0.4073)
Squared of Frequency of Meteorological Disaster Shocks	0.0356 (0.1773)	−0.1854 (0.1498)
Control Variables	Controlled	Controlled
LR chi2 (15)	43.80 ***	87.82 ***
Pseudo R2	0.2017	0.3428

Note: **, and *** indicate significance at the 5%, and 1% levels, respectively. The numbers in parentheses are standard errors.

, the impact of meteorological disaster shocks on the cooperative pest and disease control behavior of farmers with junior high school and above education is significant at the 1% statistical level, with coefficients of 1.0342. The frequency of meteorological disaster shocks also has a significant impact on the cooperative pest and disease control behavior of farmers with junior high school and above education at the 1% statistical level, with coefficients of 0.5087. However, meteorological disaster shocks and their frequency have no significant impact on the cooperative pest and disease control behavior of farmers with primary school and below education. This indicates that under the influence of meteorological disaster shocks or their frequency, farmers with junior high school and above education are more likely to engage in cooperative pest and disease control behavior. It is particularly worth noting that compared to farmers with primary school and below education, farmers with junior high school and above education have a higher likelihood of choosing cooperative pest and disease control behavior. This may be because farmers with junior high school and above education have stronger information search and processing capabilities, enabling them to quickly obtain relevant information on coping with meteorological disasters and cooperative pest and disease control behavior and understand it more deeply, adopting it more rapidly.

As shown in

Table 11. Regression results of different planting scale groups.

Variable Name	Cooperative Pest and Disease Control Behavior
	Model (3) Small-Scale	Model (4) Large-Scale
Meteorological Disaster Shocks	−0.1009 (0.4457)	0.8958 *** (0.1944)
Control Variables	Controlled	Controlled
LR chi2 (15)	28.95 **	83.40 ***
Pseudo R2	0.2369	0.2428
Variable Name	Cooperative Pest and Disease Control Behavior
	Model (7) Small-Scale	Model (8) Large-Scale
Frequency of Meteorological Disaster Shocks	−0.0743 (0.2748)	0.4558 *** (0.0980)
Control Variables	Controlled	Controlled
LR chi2 (15)	28.97 **	83.59 ***
Pseudo R2	0.2371	0.2434
Variable Name	Cooperative Pest and Disease Control Behavior
	Model (11) Small-Scale	Model (12) Large-Scale
Frequency of Meteorological Disaster Shocks	0.1830 (0.7644)	0.9107 *** (0.3188)
Squared of Frequency of Meteorological Disaster Shocks	−0.1204 (0.3427)	−0.1803 (0.1205)
Control Variables	Controlled	Controlled
LR chi2 (15)	29.10 **	85.83 ***
Pseudo R2	0.2381	0.2499

Note: **, and *** indicate significance at the 5%, and 1% levels, respectively. The numbers in parentheses are standard errors.

, the impact of meteorological disaster shocks on the cooperative pest and disease control behavior of large-scale farmers is significant at the 1% statistical level, with coefficients of 0.8958. The frequency of meteorological disaster shocks also has a significant impact on the cooperative pest and disease control behavior of large-scale farmers at the 1% statistical level, with coefficients of 0.4558. However, meteorological disaster shocks and their frequency have no significant impact on the cooperative pest and disease control behavior of small-scale farmers. A possible explanation is that farmers with larger planting scales have a higher dependence on agriculture and are more sensitive to meteorological factors. The threat of meteorological disaster shocks is greater, and once a serious loss occurs, it can significantly reduce farmers’ income. To avoid the probability of meteorological disasters and risk losses, they will be more proactive in adopting cooperative pest and disease control behavior.

5.2. Robustness Test

To verify the robustness of the benchmark model, the Winsorization method was used to conduct regression analysis on meteorological disaster shocks, the frequency of such shocks, and the squared frequency, to assess their impact on farmers’ cooperative pest and disease control behavior. In Models (1), (2), and (3) of

Table 12. Robustness test results: Winsorization method.

Variable Name	Cooperative Pest and Disease Control Behavior
	Model (1)	Model (2)	Model (3)
Meteorological Disaster Shocks	0.6786 *** (0.1695)
Frequency of Meteorological Disaster Shocks		0.3712 *** (0.0877)	0.6458 ** (0.2802)
Frequency Squared			−0.1105 (0.1074)
Decision-Maker’s Age	0.0092 (0.0087)	0.0095 (0.0087)	0.0094 (0.0088)
Decision-Maker’s Education Level	0.0243 (0.0971)	0.0124 (0.0974)	0.0245 (0.0981)
Risk Preference	−0.1340 * (0.0802)	−0.1087 (0.0815)	−0.1105 (0.0815)
Number of Family Laborers	−0.0319 (0.0658)	−0.0396 (0.0657)	−0.0373 (0.0660)
Proportion of Family Agricultural Income	0.0481 (0.2766)	−0.0295 (0.2759)	−0.0040 (0.2783)
Maize Planting Area	0.0063 (0.0048)	0.0077 (0.0047)	0.0071 (0.0048)
Maize Sales Price	0.6453 * (0.3339)	0.6463 * (0.3355)	0.6245 * (0.3364)
Type of Cultivated Land	−1.1583 *** (0.3818)	−1.2061 *** (0.3883)	−1.2094 *** (0.3847)
Government Support	0.0222 (0.1601)	0.0188 (0.1602)	0.0020 (0.1615)
Participation in Industrial Organizations	0.4238 ** (0.1708)	0.4062 ** (0.1711)	0.4088 ** (0.1715)
Social Network	0.0009 *** (0.0003)	0.0009 *** (0.0003)	0.0010 *** (0.0003)
Constant Term	−1.6651 * (0.9362)	−1.6208 * (0.9448)	−1.6416 * (0.9437)
LR chi2(15)	111.34 ***	113.01 ***	114.06 ***
Pseudo R2	0.2326	0.2361	0.2383

Note: *, **, and *** indicate significance at the 10%, 5%, and 1% levels, respectively. The numbers in parentheses are standard errors.

, the Probit model was used for regression analysis to test the impact of meteorological disaster shocks, the frequency of such shocks, and the squared frequency on cooperative pest and disease control behavior. The results show that meteorological disaster shocks and their frequency have a significant positive impact on farmers’ cooperative pest and disease control behavior. This is consistent with the previous empirical results, confirming the robustness of the research findings. Through robustness analysis, the important impact of meteorological disaster shocks and their frequency on farmers’ cooperative pest and disease control behavior is further confirmed. These results were verified under different methods and data processing approaches, indicating the reliability of the research findings.

6. Discussion

Meteorological disasters such as heavy rainfall, floods, droughts, and heatwaves have both direct and indirect impacts on the crop growth environment, thereby influencing the occurrence and spread of pests and diseases. First, these disasters can alter the environmental conditions that trigger pest and disease outbreaks. For instance, heavy rainfall and floods can waterlog fields, creating moist conditions that promote the proliferation and spread of insect pathogens and bacteria. Conversely, droughts and heatwaves can weaken crop resistance, making plants more susceptible to pest and disease attacks. Second, extreme meteorological events can disrupt the natural enemies and biological control mechanisms of pests and diseases. For example, strong winds and hail can destroy the habitats and food sources of beneficial insects, reducing their ability to control pest populations and leading to outbreaks. Additionally, these disasters can interfere with farmers’ agricultural practices and pest and disease management measures by damaging farmland, equipment, and crops, thereby weakening farmers’ capacity for effective management. Therefore, when devising pest and disease control strategies, it is essential to consider the correlation between disaster shocks and pest and disease control.

In the survey sample, only 13.21% of farmers engaged in cooperative pest and disease control behavior, indicating that a relatively small proportion of farmers participated in such activities. This form of cooperation requires coordination among farmers, involving issues such as organizational formation, information sharing, resource allocation, and mutual trust [48]. In the absence of effective organizational mechanisms, information channels, and resource support, farmers may prefer individual control measures over cooperative ones. Moreover, farmers’ knowledge and technical proficiency in pest and disease control also influence their decision-making. Insufficient knowledge and skills can make it difficult for farmers to understand and adopt unified prevention and control strategies, highlighting the importance of agricultural technology extension and training to enhance farmers’ capabilities and confidence in participating in cooperative control efforts. Finally, the social and institutional environment can affect farmers’ decisions [49,50]. Without social support, policy incentives, or institutional backing, farmers may revert to traditional individual control models. Therefore, to promote cooperative pest and disease control behavior, it is necessary to consider factors such as the institutional environment, policy support, and social incentives in a comprehensive manner.

Farmers’ level of risk perception is a crucial factor influencing their decision making and behavior regarding cooperative pest and disease control. Higher risk perception can enhance farmers’ awareness of the importance of pest and disease control. When farmers recognize the threat that pests and diseases pose to crop yields, quality, and economic benefits, they may be more inclined to adopt unified prevention and control strategies to mitigate risks. The greater the perceived risk of meteorological disasters, the more likely farmers are to cooperate and actively engage in joint pest and disease control efforts. Additionally, heightened risk perception prompts farmers to pay closer attention to pest and disease occurrences and to take preventive measures, such as regular monitoring, introducing resistant crop varieties, and using pesticides judiciously, to reduce the likelihood of outbreaks and stabilize crop yields and agricultural income. Cooperation also allows farmers to share information, resources, and experience, thereby enhancing their pest and disease control capabilities and resilience. However, it should be noted that risk perception does not always directly translate into the adoption of unified prevention and control strategies. Farmers’ decisions are also influenced by other factors, such as economic benefits, technical capabilities, social environment, and institutional support. Therefore, in addition to enhancing risk perception, it is essential to provide corresponding incentive mechanisms, technical support, and social support to encourage more farmers to participate in cooperative prevention and control efforts and ensure their sustainability and effectiveness.

7. Conclusions and Recommendations

This study investigated the impact of meteorological disaster shocks on farmers’ cooperative pest and disease control behavior, the mediating role of risk perception, and heterogeneity through Probit, IV_Probit, and mediating effect estimation models. The main conclusions are as follows: (1) Meteorological disaster shocks and their frequency significantly and positively influence farmers’ cooperative pest and disease control behavior. (2) Risk loss perception mediates the impact of meteorological disaster shocks and their frequency on farmers’ cooperative pest and disease control behavior. (3) Farmers with junior high school education or above and large-scale farmers are more likely to engage in cooperative pest and disease control behavior under the influence of meteorological disaster shocks or their frequency.

The findings suggest that the government should encourage farmers to participate in cooperative pest and disease control through measures such as financial support, technical training, information-sharing platforms, and rewards. Education and publicity efforts should be intensified to enhance farmers’ awareness of pest and disease risks and the potential losses to agricultural production. Actively establishing farmers’ organizations, cooperatives, or collaborative groups can promote cooperation and information sharing among farmers and help them collectively address pest and disease challenges. Additionally, differentiated services, incentive support, and flexible cooperation models should be provided for different types of farmers to facilitate their participation in cooperative pest and disease control activities.

Furthermore, based on the research conclusions, suggestions, and discussions, the following are the future research directions presented to the author or other relevant researchers. In many developing countries and specific agricultural regions of some developed countries, meteorological disasters occur frequently, and the agricultural production models show similar characteristics to the research samples, such as a large proportion of small-scale and decentralized farmers. Researchers and farmers in these regions can benefit greatly from the conclusions. For researchers, it provides an important framework for studying the relationship between meteorological disasters and agricultural pest and disease control. They can carry out in-depth research based on local actual situations on this basis, introduce the variable of policy support, and explore prevention and control strategies suitable for local areas. In some Southeast Asian countries where small-scale farmers mainly grow rice and frequently suffer from meteorological disasters such as heavy rain and floods, with policy support (subsidies and technical training), farmers can be actively guided to strengthen cooperation among themselves, jointly address pest and disease problems, and reduce production costs. Regardless of the agricultural production model and region, farmers’ perception of risk losses will affect their decision-making behaviors. Researchers can, based on this conclusion, conduct in-depth research on how external environmental factors can promote farmers to adopt more effective pest and disease control measures by enhancing their perception of risk losses.