When the theory of global warming was first proposed by Swedish scientist Arrhenius a century ago, few people imagined that this issue would so quickly become the most important concern of the international communityby first affecting and then changing the earth's climate (Qombar Ali et al., 2011). In other words, climate change is one of the biggest challenges that humanity is facing in the 21st century, severely affecting water resources, agriculture, energy, and tourism (Valiqolizadeh, 2018). Any relatively stable deviation in any of the climate elements or a deviation of the functioning climate factors and components from prevailing conditionswill result in climate change (Hoshmandanfar Moghadam Fard et al., 2019). Climate change can have a serious environmental, social, and economic impact on the employment and livelihood of all the planet's inhabitants, including farmers (Chisanga et al., 2017). Moreover, natural hazards such as climate change can turn into horrible and devastating accidents for human communities without risk reduction systems (Zhou et al., 2010). In this regard, societies must learn how to deal with climate change and its effects. Therefore, in the early stages, the main focus of actions was on reducing the risks caused by climate change. In the next stage, the work was directed at identifying direct responses to climate effects dealing with the resulting consequences, and adapting to these changes. Recently, the focus has shifted beyond identifying specific adaptations to resilience to ensure that individuals, communities, sectors, and nature have the necessary capacities to respond to any stress that arises (Schwarz et al., 2011).

Resilience is the third stage of response to crises and climate change (Gardner et al., 2019). The word resilience was introduced in the discussion of disaster management in 2005 at the Hugo conference. Since then, it has gradually gained a greater place in the theoretical and practical fields of disaster risk reduction. In recent years, concepts such as resilient societies, resilient livelihoods, and creating resilient societies have been commonly used in scientific articles (Manyena, 2006). Resilience as one of the types of socio-ecological systems can be interpreted in three cases: 1- the amount of disturbance that a system can absorb and remain in the same first state, 2- the limit, or the degree to which the system can organize itself and 3- the extent or degree in which the system can be built, or its capacity for learning and adaptation can be increased (Folke et al., 2004). Due to the dynamic nature of society's response to risks, resilience is a type of foresight that helps to expand policy choices in the face of uncertainty and change, which includes economic, social, environmental, and psychological dimensions (Berkes, 2007). Therefore, resilience against climate change is one of the characteristics of sustainable societies (Engle, 2011).

Climate change is a significant threat to farming communities' livelihoods (Campbell et al., 2011). The consequences of climate change and drought can lead to the instability of rural livelihoods (Speranza et al., 2008). Studies have also shown that households that depend on agriculture for livelihood are more exposed to the threats of climate change (Johnson, 2009). Therefore, rural and agricultural communities need to be more resilient against climate change because the livelihood of these communities depends on climatic conditions (Birkmann, 2011). However, the damages caused to farms and the consequences of climate change, such as periodic droughts, show farmers' lack of preparedness and resilience against this risk and its consequences. In other words, the first necessary step to deal with natural hazards such as climate change and adjust to its consequences is to know and understand the scope of people's vulnerability and their resistance to improving their tolerance and flexibility threshold (Sadeghlou and Sejasi Keidari, 2013).

Since problems and issues in rural areas differ according to the conditions of each region, it is necessary to pay special attention to regional and local planning to accurately determine the priorities in each region (Ranjbar, 2009). For this purpose, an accurate and comprehensive understanding of the existing situation, based on a systematic approach (Karimi et al., 2017). In this regard, knowledge of the spatial distribution and degree of resilience of farmers in rural areas against climate change is considered a prerequisite for development plans and programs in rural areas. Also, by evaluating and modeling the degree of resilience of farmers in rural areas against climate change using a significant share of the country's farming population, it is possible to formulate realistic plans to improve the resilience of farmers in rural areas to aid in climate change resilience efforts.

Pistachios are one of Iran's most important export products in the arid and semi-arid regions, including Kerman province (Bagheri et al., 2012). In addition to the economic aspect, pistachios play an essential rolein soil protection and sand control, especially in arid and semi-arid regions of Iran. In this province, Rafsanjan and Anar counties are considered the largest pistachio production centers in Iran, with around 80 thousand hectares of pistachio cultivation. The livelihood of most of the villagers in Rafsanjan and Anar counties is directly or indirectly dependent on pistachio production, and any disruption in pistachio production could endanger their income and, as a result, their livelihood (Hosseinifarhangi et al., 2020). As one of the main areas of pistachio production, Kerman province has faced climate change problems in recent years. Drought, cold, and untimely frosts are among the results of the phenomenon of climate change in Kerman province. Also, climate changes, drought, and indiscriminate harvesting have caused the income of the farmers to decrease (Meridenjad et al., 2014). This phenomenon has had many harmful effects on the farming and agricultural products of the region, especially pistachios, one of the primary sources of livelihood for farmers in Kerman province. During the past decade in Rafsanjan and Anar counties, the cold weather destroyed several farms and greatly reduced the productivity of others (reduced production by 70 and 85% in 2014 and 2018, respectively) (Jamalizade, 2014). Therefore, to recover their livelihood, rural households in Rafsanjan and Anar need to be able to withstand climate change and deal with crises that threaten their livelihood. Consequently, the question raised in this research is, what is the level of resilience of pistachio farmers in the rural areas of Rafsanjan and Anar cities?. And the most important dimension in measuring the resilience of pistachio farmers is the rural areas of Rafsanjan and Anar cities?.

Globally, the dominant perspective has changed from focusing only on reducing vulnerability to increasing resilience against climate change and disasters (Cutter et al., 2008). Since being introduced at the Hugo conference in 2005, resilience has become increasingly important in the theoretical and practical fields of disaster risk reduction (Mayunga, 2007). Based on Folke's definition, resilience is the capacity to absorb disruption and organization (Folke et al., 2004). Many studies on resilience opposite sciences, and the idea of resilience is an interdisciplinary concept proposed in the fields of ecology, psychology, social sciences, economics, and other sciences (Table 1).

One of the fundamental aspects of research related to resilience and resilient communities against natural hazards is finding the appropriate way to measure resilience. Since this research is related to natural hazards and disasters, researchers study ways to take steps toward practical and technological improvement to reduce the risk of accidents. Due to the multifaceted nature of resilience (including ecological, economic, institutional, and social dimensions), the transition from the theoretical framework concepts to reality is complicated and challenging to evaluate. Current resilience models examine the flexibility of societies to reduce vulnerability to the consequences of hazards; it is necessary to study and analyze these models to determine which is suitable in each case. For example, different researchers have proposed several models addressing specific aspects of resilience against flooding, while other models have been presented to measure resilience against accidents.

Many essential models are currently being used to analyze different types of resilience. Some of the most important and widely used models are mentioned below. Tobin's model has been proposed to evaluate the resilience of communities located in high-risk areas, he uses three models, risk reduction, recovery, and demographic structure, to show the sustainability and resilience of the community, and then the framework assesses which is more ecological (Tobin, 1999). Kumpfer's resilience framework is an exchange model of resilience and includes the structure of process and outcome. This model pays less attention to the cycle of destruction and reintegration and is somewhat oriented toward the interactive nature of environmental and internal content and internal resilience factors, as well as the consequences of reintegration (Kumpfer, 1999). Adger (2000) proposed a community-based disaster management model (CBDM), a bottom-up management approach that pays attention to people's participation in solving crises caused by natural disasters. Its purpose is to reduce the vulnerability of societies and strengthen people's abilities and participation to deal with the risks caused by natural disasters. The linear-time model of Davis and Izadkhah (2006) shows that society can improve its vulnerability over time in the form of a timeline in certain conditions following development. This model has three stages: 1. absorbing and tolerating stress and risk before the disaster, 2. returning to the balance after the disaster, and 3. changes in the communities to make them safe and resilient. Capital-based model (Mayunga, 2007) has been proposed as a framework for evaluating society's resilience against disasters based on types of capital (material, economic, physical, human, and natural). The spatial model or DROP (Cutter et al., 2008) is designed to clarify the relationship between resilience and vulnerability and provides a comparative assessment of disaster resilience at the local and community levels. This model defines resilience as a dynamic process dependent on previous conditions, the severity of accidents, the time between hazards, and the influence of exogenous factors. This model pays attention to the economic, social, ecological, and institutional dimensions. Jurjonas and Seekamp's RCCR model 2017 considers resilience as a range from vulnerable to flexible, where the capacity to adapt to society is scaled between opposite indicators. The drawn line shows the relative position of societies in this spectrum or at what level the whole system is placed.

Various studies have been proposed and conducted to measure the resilience of farmers in rural areas sample of domestic and foreign studies in this field is presented in Table 2.

After considering the different dimensions of resilience and the multiple definitions of this concept, this study used the RCCR model to measure resilience. The RCCR framework is designed to rapidly engage stakeholders with other stakeholders to assess local resilience needs (Jurjonas and Seekamp, 2017). Furthermore, this model considers resilience as a range from vulnerable to flexible, where the capacity of a society is scaled between opposite indicators. Therefore, this model aims to reduce the vulnerability of societies and strengthen people's abilities and participation to deal with the risks caused by natural disasters and accidents. In other words, RCCR integrates the idea that adverse changes affect societies, and high vulnerability can eventually lead to the collapse of the whole system (Jurjonas and Seekamp, 2017). Table 3 presents the conceptualization of the dimensions of the RCCR model.

The current research was conducted as a survey and used quantitative numerical data analysis in a descriptive-correlation method. The statistical population of the research was 27,860 pistachio farmers in rural areas of Rafsanjan and Anar counties (Figure 1). A stratified sampling method was used to select the studied villages with regard to the extent of the studied. Each village was first divided into four classes (north, south, west, and east). Villages were then randomly selected in each class (map 2). Next, Cochran's formula was used to determine the sample size of 588 users among the 27,860 users (pistachio farmers) in the study area, with an error level of 4%. Lastly, several beneficiaries were randomly selected in each village, a form of proportional assignment.

To collect the research data according to the research objectives, a multi-section questionnaire was designed with questions using a qualitative range of five options (completely disagree = 1, disagree = 2, have no opinion = 3, agree = 4, and completely agree = 5). The data collection process lasted three months, from May-July 2018, and data was collected from 60 villages in Rafsanjan and Anar. The RCCR resilience framework consisting of five components, livelihood diversity, wealth, sustainable development, social cohesion, and flexibility, was used to measure the structure of resilience against climate change (Jurjonas and Seekamp, 2017). The operational definitions of the variables used in this research are presented in Table 4.

The apparent validity method, using the opinions of professors and experts in several stages of modification and revision, was used to determine the validity of the questionnaire. To determine the reliability of the questionnaire, Cronbach's alpha coefficient for a pre-test (30 questionnaires) and the study questionnaire was calculated, 0.914 and 0.682, respectively, showing the validity of the designed questionnaire. To model the resilience behavior of the farmers in the studied rural areas against climate change (spatial analysis), the degree of resilience against climate change was first measured. From several methods presented for measurement and evaluation (such as numerical taxonomy, TOPSIS, weighted simple sum, Vaikor, etc.), the Vaikor method was found to be appropriate for this study (Karimi et al., 2017). Next, the entropy method was used to weigh the studied indicators. Then, the results obtained from the combined Vaikor-entropy model (in the range of zero and one) were entered as the input of the software of the Geography Information System and with the help of interpolation and the inverse weighted distance method. Finally, Inverse Distance Weighted (IDW) resilience behavior against climate changes was modeled in raster form. Data processing in this research was done using MATLAB 7.10, AMOSver24, SPSSver25, and ArcGIS 10.5 software.The utilized research methods are explained in more detail in the following paragraphs.

Vaikor is an applied multi-criteria decision-making method with high efficiency in solving discrete problems. This method is based on consensual planning, and in it, the consensual solution determines the justified solutions that are close to the ideal solution and is created in the form of an agreement through special credits of the decision-makers (Opricovic and Tzeng, 2004). In this method, the emphasis is on ranking and choosing from a set of options and determining a consensus solution for a problem with conflicting criteria (Chen and Wang, 2009). Presented by Buyukozkan and Ruan (2008) and by Opricovic and Tzeng (2004), the basis of consensual models was developed based on the LP metric method. The following steps are implemented to use this method (Opricovic and Tzeng, 2004).

Step 1: Formation of the spatial decision matrix.

Step 2: Determine the best and worst value for all criteria functions. If the criterion function indicates profit (positive) and cost (negative), the best and worst values are calculated based on relation 1.

Step 3: Determine the weight of the indicators. In this field, many methods, such as rank-order analysis, entropy, eigenvector, etc., can be used according to the need. The entropy method is used in this research.

Step 4: Calculate the values of the distance between the options and the ideal solution. In this step, the distance of each option from the positive ideal solution is calculated, and the calculated weights are included; then, its aggregation is calculated based on the following formulas.

where f_ij is the index I in unit j, f_jth is the positive ideal of index I f_j- is the negative ideal of index I, W_i is the weight of the i-th index obtained from the fourth step, Sj is the distance from option I to the ideal solution, and Rj is the distance of option I from the negative ideal solution (worst combination)

Step 5: Calculate the value of Qi and Kor for i=1, 2, ..., m; The value of Qi is calculated based on the following equation.

Step 5: Calculate the value of Qi and Kor for i=1, 2, ..., m; The value of Qi is calculated based on the following equation. Qi=v[Si-S*S--S*]+(1-v)[Ri-R*R--R* ]

where V is the weight of the strategy (the majority of criteria) or the maximum group favorability. (S_i-S^∧*)/(S^∧–S^∧*) shows the distance from the positive ideal solution of the i-th option. In other words, (R_i-R^∧*)/(R^∧–R^∧*) indicates the distance from the negative ideal solution for the i-th option. When v is >0.5, the Qi index has the maximum agreement. When V is smaller than 0.5, it indicates the maximum negative attitude.

Step 6: Ranking the options based on Qi values. Based on the values of Qicalculated in the fifth step for the options, the desired value can be reduced from one; thus, the village with the highest score (Qi-1) has priority in the selection.

In information theory, entropy is a measure of uncertainty expressed by a certain probability distribution Pi (Hwang et al., 1981). The steps of the entropy method are described below (Karimi, 2013).

The first step is the calculation of normalized data: equation 4 is used for normalization.

The second step is to calculate the entropy of the factor j (E_j) for E_j from the set of P_ij for each index using equation 5.

In the third step, the value of the degree of deviation, (dj)¹, is calculated, showing how much useful information the jam index provides to the decision maker, and is expressed in the form of equation 6:

The fourth step is to calculate the weight of existing indicators and factors (Wj):

The confirmatory factor analysis results showed that the path coefficients between the indicators and the hidden variable are significant, and the CR indexes and the average variance extracted for all these variables have high and appropriate values. Therefore, it can be stated that all the indicators selected to measure the latent variable of resilience against climate change were correctly selected, and their validity and reliability were also confirmed. In addition, the dimension of livelihood diversity had the highest standard coefficient; therefore, it can be said to be the most important dimension in measuring the resilience of pistachio farmers in the rural areas of Rafsanjan and Anar counties against climate change. Also, the results obtained from measuring the degree of resilience of pistachio farmers in rural Rafsanjan and Anar areas against climate change show that the current situation of the farmers in this field is at a medium to a low level. In the meantime, the village of Gurban-Abad in the Hossein-Abad sub-district had the highest degree of resilience, and the village of Deh Abbas in the sub-district of Khnaman sub-district had the lowest level of resilience. The results also indicated that the status of the indicators in the livelihood diversity dimension is in a more suitable situation than other dimensions. Moreover, the status of the indicators of the sustainable development dimension is in an unfavorable situation compared to other dimensions. Also, after analysis, the studied villages were divided into six levels, ranging from lack of endurance to resilient. None of the villages fell into the lowest or highest level. There are 8 villages at the very weak resilience level and 22 villages at the weak resilience level. There are 26 villages at the relatively resistant level and four villages at the level of good resilience. According to the successful results obtained from this model, this model can be used for future studies in other countries and provinces. This model can also be used to compare a product between two provinces with the same parameters.

In recent years, climate change and drought as natural phenomena have created many challenges in different regions of Iran. Coping with climate change is important from different aspects, one of which is related to the performance of managers and another related to farmers. Therefore, an approach that will strengthen the resilience capabilities of farmers and also reduce their vulnerability will increase the resilience of rural settlements. However, the results showed that the resilience status of the farmers in the studied area is lower than the average (64.97% at the average to low level). This finding indicates that the studied pistachio farmersare not in good condition in terms of resilience and are unable to deal with wider climate changes or any fluctuations in climate changes (reduction of rainfall, frost, etc.) This inability could cause dire consequences in the agricultural sector of the region. Thus, plans to make the users more resilient against the effects of climate change and increase the tolerance threshold in the Rossinitai area should be prioritized.

It should be noted that the results obtained in the field of measuring resilience against climate change are in line with the findings of Bagheri Fahroji et al., 2018.

In the end, according to the findings of the research, suggestions are made to improve the resilient behavior of farmers against climate change.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

Study conception and design: AA, MO, and JF. Data collection and analysis and interpretation of results: AA and MO. Draft manuscript preparation: AA. All authors reviewed the results and approved the final version of the manuscript.