Risks and farmers' behavior change towards water conservation: a study in the Southeast of Iran

In the contemporary world, issues related to the management of water resources, e.g., droughts and water scarcity, have posed increasing challenges for crop production and human life (Mittal et al. 2016; Que et al. 2022; Karimi & Ataei 2023). These factors, along with economic growth and expansion of agricultural lands, have put much pressure on water resources (Javeline et al. 2019; Patra et al. 2023), whereas changes in water resources are remarkably influenced by climate change and human activities. Compared to the effects of climate on water cycles and availability, human activities play a more important role in controlling water quality and quantity (Adams 2014; Woldesenbet et al. 2017; Hu & Zhang 2022). Recent droughts are accompanied by increased demand, population growth, and agriculture development and have triggered water scarcity in different parts of Iran (Yazdanpanah et al. 2014; Yang et al. 2022). In agrarian societies, the effects of water scarcity are manifested in socioeconomic aspects (Keshavarz et al. 2013; Maleksaeidi & Karami 2013; Panyasing et al. 2022). The unwise decisions of governments and officials in charge of resource management (Walter et al. 2022; Parwada & Marufu 2023) and inattention to the consumption behaviors of people, especially farmers, have created a lot of problems in the use of water resources. Due to the overuse of water resources, 307 out of 609 plains in Iran have been declared critically endangered, as the groundwater levels or aquifer quality are constantly dropping (Hashmi & Al-Modarressi 2015). The Sistan plain in the southeast of Iran is a relatively vast low area composed of old and new alluviums of the Hirmand, Fararoud, Khasroud, and Nahbandan rivers. A part of this plain is a desert that turns into a wetland, one of the most extensive wetlands in the south of Iran, by hosting a massive volume of flood water annually (Azadeh et al. 2010; Karimi & Ataei 2022). The most important river in the region is the Hirmand, which originates from the Hindu Kush Mountains in Afghanistan. The Sistan region is struggling with the most intensive drought in Iran, which originates from the sharp decline in precipitation and its complete reliance on the water supplied by the Hirmand River, while the river is dry most of the time. This region, as an agricultural hub, heavily relies on surface water resources, and the majority of its population lives in rural areas (Mokhtary & Iraj 2008; Hallaj et al. 2021).

Research (Bijani & Hayati 2015; Nguyen-Chi et al. 2022; Bhakta et al. 2023) shows that an average of 70% of water resources are consumed by the agricultural sector (Yazdanpanah et al. 2014; Mamani et al. 2022; Vasilaky et al. 2023; Zuo et al. 2023). So, the sound and optimal use of water resources is an urgent necessity in the agricultural sector. Water conservation is the primary strategy in future water planning and management and reflects pro-environmental activities (Gilbertson et al. 2011; Adams 2014; Nayar & Patel 2021). It is also the most critical measure to protect water resources by which the interests of future generations can be secured and the demand for sensitive water-dependent ecosystems, such as lakes and rivers, can be reduced. Water conservation includes behaviors for reducing water consumption (Tajri Moghadam et al. 2019), and these behaviors are promoted when farmers understand the risks and hazards of water scarcity (Sharafipour & Ahmadvand 2019). Furthermore, the perceived efficiency of the response or the perceived effectiveness of the conservation measures for mitigating risks will play a key role in water conservation behavior. Also, people's belief in their capability or farmers' perceived self-efficiency, as well as the perceived costs of conservation measures, will influence their behavior (Sharafipour & Ahmadvand 2019). It is very important to identify farmers' behavior and actions towards water resources and their conservation. In the environmental behavior literature, there is a wide range of studies on behaviors, and each has somehow concentrated on analyzing the impacts of human behaviors on the natural environment. Environmental psychology is a field of study that deals with the interactions between human and their environment, the psychological roots of environmental degradation, and the relationship between environmental attitudes and pro-environmental behaviors such as water conservation behavior (Kollmuss & Agyeman 2002; Aliabadi et al. 2022; Ataei et al. 2022). The use of behavioral science for understanding water conservation and the use of psychological and anthropological theories and models for understanding behavior-predicting variables are such that their understanding brings about behavior modification (Yazdanpanah et al. 2015; Khoshnodifar et al. 2023).

So, this research aims to explore farmers' behavioral processes for water conservation in southeast Iran. The novelty of the research is that few studies have addressed the effect of farmers' perceived risk and risk attitudes on their behavioral processes in the context of water conservation. This research can provide a deep and new understanding of farmers' mentality about risk management and their water conservation behavior. This study attempts to fill the gap in the literature on behavior changes by considering farmers' perceived risk and risk attitudes and different dimensions of the health belief model (HBM). Previous studies have mostly dealt with farmers' social dimensions and have less focused on risk dimensions of climate changes. As such, this research can supplement the literature from social, cultural, and psychological perspectives.

The HBM was developed by social psychologists Godfrey H. Hochbaum and Irwin M. Rosenstock in the 1950s (Strecher & Rosenstock 1997). HBM is the most common and popular theoretical model for promoting preventive health behaviors and is used to describe health-related behavior change or persistence (Carley & Stratman 2015). HBM focuses on people's beliefs about their decisions (Tarkang & Zotor 2015) and emphasizes how a person's perception creates motivation and movement and changes their behavior. Generally, HBM focuses on changing beliefs and states that a change in belief triggers a change in behavior (Namdar et al. 2012; Yazdanpanah & Mardasi 2017). Other advantages of HBM are that it has a theoretical framework to better understand and create stimuli for the cooperation of deprived people in health education programs and that it reveals the factors that are effective in encouraging or discouraging an individual for engaging in certain health-related measures (Cook 2018). It should, however, be noted that a review of the literature on HBM shows the focus on interpersonal factors, whereas some external factors, e.g., education and economic factors, influence intra-personal factors. Factors like the level of knowledge, attitude, beliefs, motivation, and so on are promoted by education and lead to better decision-making (Vermandere et al. 2016; Dodel & Mesch 2017).

HBM predicts two main groups of behavioral beliefs: perceived threats and perceived expectations. Perceived threat is composed of two sub-components: perceived susceptibility (PSu) and perceived severity (PS). Perceived susceptibility refers to people's perception of a hazard or the chance of suffering from health problems. Perceived severity is defined as the extent to which an individual judges the medical and social consequences of the problem (Witte 1992). Perceived benefits (PBe) refer to the effectiveness of certain actions to alleviate the hazard and preserve health, perceived expectations refer to positive results of sound actions, and perceived barriers (PB) refer to the barriers that, in an individual's opinion, prevent the implementation of healthy behavior (Zetu et al. 2014).

Researchers argue that the structure of HBM is based on constructs that depend on behavior persistence, such as perceived severity, perceived susceptibility, perceived benefits, perceived barriers, self-efficacy, cues to action (CA), and general beliefs. Also, self-efficacy, cues to action, and general beliefs are factors added by researchers to HBM because they argue that these new components can promote the descriptive power of the model (Tashiro 2022). Cues to action encompass a range of incentives, e.g., physical events, social media, and social impact, which activate an individual's readiness by changing through awareness of the negative consequences of unhealthiness (Simsekoglu & Lajunen 2008; Tengecha et al. 2022). Self-efficacy is an individual's perceived ability to perform an activity, and general beliefs represent personal values, specific beliefs, and concerns about health problems (Li et al. 2015). This model assesses the relationship between health-related beliefs and preventive healthy behaviors (Razmara et al. 2018). According to this model, if people feel exposed to a situation (perceived sensitivity), they will engage in preventive actions to hinder the hazard. Furthermore, an individual will most likely engage in the required behavior if they believe that the situation is potentially dangerous and can have significant effects (perceived severity) and that the hazards and side-effects of the situation can be reduced by some measures whose benefits (perceived benefits) outweigh the barriers on the way of adaptation to the behavior (e.g., perceived time and monetary barriers) (Yazdanpanah et al. 2015).

The HBM attempts to predict behaviors by accounting for individual differences in beliefs and attitudes. However, it does not account for other factors that influence behaviors. For instance, habitual behaviors may become relatively independent of conscious decision-making processes (Siddiqui et al. 2016). Environmental factors outside an individual's control may prevent engagement in desired behaviors (Glanz et al. 2008). Furthermore, the HBM does not consider the impact of emotions on behavior. Another limitation of the HBM is that factors other than beliefs also heavily influence behavior practices. These factors may include special influences, cultural factors, socioeconomic status, and previous experiences (Orji et al. 2012). Therefore, this research tried to extend the HBM by adding new components: risk attitude and perceived risks. The advantage of HBM is that it well explains the relationship between people's beliefs and their intentional behaviors. Accordingly, if an individual is exposed to a hazard, they will intend to take preventive behaviors provided that the hazard has extensive negative consequences for the individual and the preventive behaviors can effectively alleviate its damages or severity (Gaines & Turner 2009).

Farmers' perceived risk and risk attitudes toward it are among the factors influencing the adoption of risk management tools for water conservation (Sánchez-Cañizares et al. 2022) and have a significant impact on their decision based on the strategies adopted for risk management (Rahman et al. 2022). In other words, perceived risks and risk attitudes are vital psychological factors that influence water conservation behavior (Wang & Liu 2021). The perception of risk varies among social communities and is mostly shaped by subjective feelings towards various activities, events, and technologies (Wang & Liu 2021). People's perception of risk plays a key role in their decision-making process (Salehi et al. 2018). Risk perception guides decision-making regarding the acceptance of risks and fundamentally affects behaviors before, during, and after a disaster (Marshall 2020). As well, perceived risk is a subjective assessment of the likelihood of the occurrence of a certain incident and how to be worried about its consequences (Fierros-González & López-Feldman 2021). Therefore, climate change is intensifying, and local communities, particularly farmers, must adapt and cope with its effects. They also need to conserve water resources for their livelihood sustainability. Thus, we need to understand the risk aspects of farming in these situations and study how perceived risks affect decision-making behavior when investigating people's choices about climate change (Karimi et al. 2018). Farmers' perception of risk and management strategies to prevent risk, as well as their socioeconomic characteristics, is important factors in predicting water conservation behavior (Nadhomi et al. 2013).

Many research studies have studied water conservation by using psychological-cognitive models. Corral-Verdugo et al. (2003) studied environmental beliefs and water conservation and modeled general environmental beliefs as a three-factor structure composed of beliefs in (1) the need for protecting balance with nature, (2) the need for limiting human growth, and (3) a human exception model. Based on their results, general environmental beliefs influence the development of specific beliefs about water differently. Water beliefs are positively influenced by this model, whereas ecological water beliefs are positively influenced by beliefs in limitations and are negatively related to the human exception model. Water beliefs, in turn, increase water consumption, whereas their ecological water beliefs restrain their behavior. Bayard & Jolly (2007) adopted the HBM to study the effects of perceived susceptibility, perceived severity, and perceived benefits and barriers on farmers' awareness of and attitude to environmental behavior and environmental degradation. According to the results, perceived benefits of land improvement positively influence farmers' attitudes at a higher economic level. Also, the perceived severity of land erosion positively influences their awareness and attitude.

In a study on self-efficacy and its effect on water conservation behavior, Sharafipour & Ahmadvand (2019) revealed the significant effect of self-efficacy and its non-costliness on water conservation and management behavior. Tajeri Moghadam et al. (2020) investigated the barriers to water conservation behavior with grounded theory. They listed wrong beliefs and habits, fatalism, and the lack of understanding and responsibility about water scarcity as the causal conditions, land defragmentation, economic and financial limitations, poor governmental planning and management, lack of cooperation and collective consensus, lack of cultural development and training, and distrust to experts and officials as the contextual factors, and farmers' lack of self-efficacy, poor support plans of the government, and the failure in supplying farmers' livelihoods and benefits as the barriers to conservative behavior. Tajeri Moghadam et al. used the HBM to study water conservation behavior and found that perceived susceptibility, perceived benefits, and cues to action accounted for 41% of the variance in farmers' water conservation behavior. Furthermore, perceived benefits were found to be the strongest predictor of water conservation behavior. They concluded that the focus must be on perceived benefits, perceived susceptibility, and cues to action to promote preventive measures in coping with the growing water scarcity in Iran.

Moradhaseli et al. (2021) explored the underlying factors of farmers' occupational health behavior using the HBM and revealed that their behavior was influenced by four dimensions: perceived susceptibility, perceived benefits, cues to action, and perceived self-efficacy. These findings confirm the potential of the HBM in describing farmers' behaviors and the relevance of the model's use in studying farmers' protective behaviors. Ataei et al. (2021) employed the theory of planned behavior and HBM to investigate farmers' behaviors toward the use of green pesticides. They stated that among the seven constructs of HBM, the four constructs of perceived susceptibility, perceived benefits, cues to action, and health motivation were significantly related to the intention to use green pesticides.

The literature review shows that water conservation behavior is a critical issue investigated by many researchers. However, perceived risks of drought and farmers' risk attitude are two vital components whose role in exhibiting a behavior has been underestimated by researchers. Notably, if farmers have no perception of the risks of drought and water scarcity and have a poor risk attitude toward drought, their water conservation behavior process may change. Accordingly, there is a theoretical and practical gap in the research on water conservation behavior, which needs further research. In an attempt to understand the process of exhibiting water conservation behavior, this research adopts the HBM and includes farmers' risk attitudes and perceived risks in the model. The information collected greatly contributes to understanding how farmers manage water scarcity risks and measuring their knowledge of and attitudes toward water risks and advantages. Despite the importance of farmers' perceived risk of water scarcity and water conservation, no research has ever used the HBM to investigate farmers' water conservation behavior in the Sistan plain.

This study offers several noteworthy contributions to the existing literature. Firstly, it identifies the farmers' risk level in drought. This detailed analysis provides valuable insights into the farmers' decision-making process towards water conservation, contributing to a deeper understanding of the implications for drought mitigation and adaptation. Secondly, the study investigates the components of HBM influencing farmers' behavior. This insight provides valuable guidance for designing effective intervention strategies and promoting behavioral changes among farmers, ultimately facilitating the adoption of water conservation practices. In conclusion, this study bridges a crucial research gap in drought adaptation by evaluating the risk factors and exploring the underlying factors driving farmers' behavior.

The results for all constructs (α = 0.73–0.94; CR = 0.78–0.89) implied their proper consistency. The variables were all calculated on a 5-point scale from 1 for ‘completely disagree’ to 5 for ‘completely agree’ and from 0 for ‘none’ to 5 for ‘very much.’ The data were analyzed using mean, percentage, and structural equation modeling in the SPSS_win26 (Field 2017) and AMOS₂₄ (Byrne 2016) software suites. The indices of chi-square value/degree of freedom (CMIN/DF) (less than 5), root mean square residual (RMR) (greater than 0.90), Adjusted Goodness of Fit Index (AGFI) (greater than 0.90), comparative fit index (CFI) (greater than 0.90), goodness of fit (GFI) (greater than 0.90), and Root Mean Square Error of Approximation (RMSEA) (less than 0.08) were used to investigate the structural model's goodness of fit (Byrne 2016).

According to the results for the farmers' demographic characteristics, 93.9% were male and 6.1% were female. The farmers of the Sistan plain were, on average, 51.18 years old (with a standard deviation (SD) of 10.77), and they have been working in the field of agriculture for, on average, 29.92 years (SD = 11.64). Most farmers were literate at the basic and middle-school levels (20.5 and 19.4%, respectively). Furthermore, 12.7% were single and 87.3% were married. Regarding land ownership type, most farmers (72.6%) were the owners of their lands, whereas 27.4% were tenants. The mean cultivation area of agricultural and horticultural crops was 13.46 ha (SD = 3.13). The river was the main water resource for most farmers (61.8%), but 30.2% abstracted groundwater tables through water wells. Also, the main sources of the farmers' revenues were horticulture (57.1%), crop farming (36.6%), and animal farming (6.4%) (Table 1).

Confirmatory factor analysis (CFA) is the fundamental first step in running most types of structural equation modeling (SEM) models. It is used to verify the measurement quality of any and all latent constructs used in the structural equation model. The first step is to calculate the factor loadings of the indicators (standardized loadings) that make up the latent construct. The standardized factor loading squared is the estimate of the amount of the variance of the indicator that is accounted for by the latent construct (Cheung et al. 2023). Creating this CFA measurement model allows us to check the convergent validity of the construct. Convergent validity is indicated by high indicator loadings, which shows the strength of how well the indicators are theoretically similar.

Table 2 presents the factor loadings of each studied indicator on its corresponding construct. Accordingly, the variable of water conservation behavior is composed of nine indicators. The t-value was greater than 1.96 for all of them. So, the factor loadings of the indicators of this variable significantly differed from zero. Also, all t-values were greater than 1.96 for the indicators loaded on the variables of perceived risks, risk attitude, perceived benefits (PBe), perceived barriers (PB), perceived severity (PS), perceived susceptibility (PSu), perceived self-efficacy, and cues to action (CA), implying that the factor loadings corresponding to these indicators were their significant predictors. The research also computed composite reliability (CR) and average variance extracted (AVE). The analysis of these indicators showed that CR was greater than 0.6 for all variables and AVE was greater than 0.5 for all of them.

The structural model's goodness of fit was found to be acceptable based on the fit indices, including CMIN/DF, RMR, AGFI, CFI, GFI, and RMSEA. Furthermore, GFI, an indicator of the relative size of variances and covariances captured by the model, was estimated at 0.91 for the research framework. The CFI was 0.92, meeting the optimal level like other fit indices. χ²/degrees of freedom was calculated to be 3.25, indicating the proper fit of the model. Also, the values 0.90, 0.94, and 0.06 were calculated for RMSE, AGFI, and RMR, respectively. Given the fit of the structural model of the study, it can be said that this model is generally consistent with the data used.

The results revealed that perceived benefits (β = 0.42, P < 0.05), perceived barriers (β = 0.62, P < 0.01), perceived severity (β = 0.78, P < 0.05), perceived susceptibility (β = 0.62, P < 0.01), cues to action (β = 0.42, P < 0.01), and perceived self-efficacy (β = 0.61, P < 0.01) influenced the farmers' water conservation behavior positively and significantly. So, hypotheses 1–6 are supported.

Also, the farmers' risk attitude was found to influence perceived benefits (β = 0.38, P < 0.05), perceived barriers (β = 0.19, P < 0.05), perceived severity (β = 0.25, P < 0.05), and perceived susceptibility (β = 0.37, P < 0.01) positively and significantly. This confirms hypotheses 8–11. However, hypothesis 7 is refuted because the farmers' risk attitude had no significant effect on perceived self-efficacy.

The positive and significant effect of the farmers' perceived risks was revealed on perceived severity (β = 0.28, P < 0.05), perceived self-efficacy (β = 0.30, P < 0.01), and perceived barriers (β = 0.23, P < 0.05). Accordingly, hypotheses 12, 14, and 15 are supported. The farmers' perceived risks, however, had no significant influence on perceived susceptibility and perceived barriers, refuting hypotheses 13 and 16 (Figure 2).

R-squared (R²) is a statistical measure that represents the proportion of the variance for a dependent variable explained by an independent variable in a regression model. R² was found to be 0.15, 0.13, 0.287, 0.133, and 0.08 for perceived benefits, perceived barriers, perceived severity, perceived susceptibility, and perceived self-efficacy, respectively. This means that 15, 13, 28.7, 13.3, and 8% of the variances in these variables are related to the farmers' perceived risks and risk attitudes. Also, R² was estimated at 0.511 for the farmers' water conservation behavior, meaning that 51.1% of the variance in farmers' water conservation behavior is predicted by perceived benefits, perceived barriers, perceived severity, perceived susceptibility, cues to action, and perceived self-efficacy (Table 3).

This research integrated risk attitude and perceived risks with the components of the HBM to investigate farmers' water conservation behavior. The results proved that the extended HBM was highly successful in predicting farmers' water conservation behavior. This finding has been supported by many research studies in different fields (Li et al. 2015; Dodel & Mesch 2017; Cook 2018; Jeong & Ham 2018; Ataei et al. 2021). Our results agree with some previous studies on the components of perceived barriers and benefits. Abdollahzadeh & Sharifzadeh (2021) and Abazari et al. (2022) established that understanding the barriers and benefits of action would increase the likelihood of its repetition. Some benefits of using water conservation strategies include increasing the water level of wells, enhancing water productivity, escalating production, and reducing production costs.

According to the results, Perceived susceptibility is a strong predictor of water conservation behavior. This has been corroborated by other studies (Ataei et al. 2021; Renault et al. 2021; Sereenonchai & Arunrat 2022; Tengecha et al. 2022). According to these scholars, perceived susceptibility is a determining factor in people's behavior. Farmers who believe that they are exposed to the risk of drought and water scarcity hinder the occurrence of drought effects or take protective measures to adapt to and alleviate its impacts. So, water conservation behavior is reinforced among them. Also, it was found that cues to action would improve the farmers' behavior towards water conservation. This is in agreement with the results of Teshome et al. (2021), Win et al. (2021), and Yazdanpanah et al. (2022), who found that educational messages, collective communications, entrusted people, and entrusted mass media can influence people's behavior. In other words, human and communication accelerating factors induce the need for people to take water conservation measures. So, their measures for fructifying conservation behavior are reinforced. This implies that individuals are affected by knowledge sources. For farmers' water conservation behavior, CA can change the way towards displaying proper behaviors.

PSe is a construct that significantly influences the process of displaying water conservation behavior by farmers. This finding is related to the results of Zobeidi et al. (2021), Tashiro (2022), and Abazari et al. (2022). It can be asserted that farmers' understanding of their ability to use water conservation strategies can improve their conservation behavior. In other words, water conservation behavior needs farmers' full understanding of their abilities, which can speed up the process of behavior exhibition. As well, the results reveal that risk attitude and perceived risks affect most components of the HBM significantly. Therefore, farmers who perceive greater risk, either due to the fact that they are objectively faced with greater risk or they have a greater subjective perception of various risks, have a significantly greater perception of benefits and barriers of the use of water conservation strategies, drought susceptibilities and severity, and self-efficacy. Hellerstein et al. (2013), de Mey et al. (2014), and van Winsen et al. (2016) have emphasized the role of risk in farmers' agronomic activities, too.

Water scarcity is on the list of environmental challenges and an important factor jeopardizing sustainable socioeconomic development in the world, especially in arid and semi-arid regions like Iran. One way to alleviate the consequences of this global crisis is to change farmers' behavior toward the conservation of water resources. Accordingly, we developed a theoretical model for understanding water conservation behavior based on the HBM and the intervention of risk attitude and perceived risks. Empirical evidence for this model was collected with a survey of farmers in the Sistan region. This model integrates perceived risks and risk attitude as the factors determining the HBM components, whereas water conservation behavior is affected by perceived susceptibility, perceived severity, perceived benefits, perceived barriers, self-efficacy, and cues to action. In conclusion, this study fills a significant research gap by evaluating the risk dimensions of climate change and investigating farmers' water conservation behavior.

The results showed that the theoretical framework of the study was successful in explaining farmers' water conservation behavior. So, the present study supports the theoretical framework as a strong predictor of behavior. Overall, the results reveal that to promote water conservation behavior and use water scarcity-coping strategies in Iran, it is imperative to focus on components like perceived benefits, perceived susceptibility, perceived severity, perceived barriers, cues to action, and self-efficacy while farmers' understanding of the risk will facilitate this process. To motivate farmers to adopt drought adaptation and coping strategies, it is recommended to enhance their awareness of the negative impacts of the degradation of water resources through governmental policy programs on the one hand and to consider the benefits of adopting water scarcity-coping strategies on the other.

As with all studies, this research faced some limitations. One limitation was the difficulty in detecting farmers who applied water conservation practices. On the other hand, it was difficult to gain their trust to complete the questionnaire. To solve this problem, they were informed about the research goals. Also, they were ensured about the confidentiality of their responses and that the data would not be revealed to other organizations. Another limitation was to study farmers' water conservation behavior in detail. In other words, it is impossible to determine the level of the farmers' water conservation behavior, and there is no comprehensive standard to show whether their behavior is optimal. This can be subject matter for future research to develop a framework or guideline for farmers' water conservation behavior level. Another limitation was farmers' self-reporting of the application of water conservation measures. In other words, the metric for assessing their behavior was their self-reports, which can be changed in future research.

Data cannot be made publicly available; readers should contact the corresponding author for details.

The authors declare there is no conflict.