Farmers' resilience index: A tool to metricize the resilience of the farmers towards natural disasters affecting agriculture in India

India is vulnerable to a wide range of both natural and man-made disasters. Approximately 58.6% of the continent is vulnerable to earthquakes of moderate to very high intensity, floods and river erosion, while nearly 5,700 km of the 7,516 km long coastline is vulnerable to cyclones and tsunamis. Drought is a risk for more than 68% of the cultivable land (National Disaster Management Authority, 2019). However, the eastern coast of India is more vulnerable to disasters than the west. Between 1891 and 2000, the East Coast of India was hit by over 308 cyclones, of which 103 were classified as severe. Over 62 cyclones, including depressions, surges and severe surges, have hit the Andhra Pradesh coast in the last 40 years. There were 32 cyclones that hit the Krishna–Godavari area, which includes the four districts of East Godavari, West Godavari, Krishna and Guntur (Basheer Ahammed & Pandey, 2019).

The recent Gulab cyclone tragedy that occurred in 2021 caused extensive damage to agricultural fields in the coastal states, including Telangana, Odisha and Andhra Pradesh due to cyclone storms and heavy precipitation. At the same time, the flood-affected croplands in Andhra Pradesh were found in several regions (Prakash et al., 2023). Regular cyclones in India result in a significant reversal of the country's developmental progress since they cause a high number of fatalities, lost livelihood chances, loss of both public and private property and severe infrastructural damage.

Andhra Pradesh is a coastal state in southeastern India. It is one of the most exposed states in India to cyclones. Its proximity to the Bay of Bengal and its low-lying topography are compounded by high population density, and poorly maintained flood protection and drainage systems making coastal Andhra Pradesh highly vulnerable to the impacts of cyclones and floods (Rao et al., 2017).

Agriculture is the main and only source of income for a majority of the farmers in Andhra Pradesh, especially in rural areas. It is one of the most disaster-affected sectors. Even though the state is referred to as the ‘rice bowl of India,’ there are problems with the state's agriculture. Certain areas – particularly the coastal districts – are more likely to experience cyclones and floods, whereas the Rayalaseema regions are more likely to experience persistent drought (NADMP, 2020). Though it is not possible to prevent natural disasters, it certainly is possible to reduce the frequency and intensity of the disasters by evolving appropriate preparedness to counter disaster and initiate mitigation measures.

The Indian Government has set a target of doubling farmers' income by the year 2022, 5 trillion-dollar economy and ease of living by 2024. However, these targets can be achieved by improving the scenario of Indian agriculture and protecting them from natural disasters through adaptation. According to a recent research by the Ministry (2018), if suitable adaptation measures are not implemented in the agriculture sector, climate change is predicted to have a more severe impact on agricultural productivity starting in 2020 and might increase by as much as 40% by 2100 (NADMP, 2020).

The concept of resilience stems from the Latin ‘resilire’ to denote ‘bouncing back’ or ‘recoiling’. The IPCC (2007) defined resilience as ‘the capacity of a social or ecological system to absorb disturbances while preserving the same fundamental structure and modes of functioning, the capacity for self-organization, and the capacity to adapt to stress and change’. For agricultural production systems to be successful, farmers’ capacity to absorb, handle, and recover from a climate disaster is crucial. At the village level, farmer resilience is all about giving them the tools they need to absorb and recover from shocks and pressures that affect their agricultural output and way of life. When anticipating a crisis in the near future, communities can learn from their proactive responses to climate disasters.

As resilience is a dynamic multidimensional concept, its quantification remains controversial (Béné et al., 2016). Measuring resilience is a difficult task and different authors proposed different approaches. Alternative measurement methods or resilience indicators are often used since measuring resilience is not easily possible (Quandt, 2018). The climate disaster resilience (CDR) framework adopted to study the farmers' resilience index which showed that the economic dimension is high (0.58) when compared with the social, technical and physical dimensions which have farmers' resilience index value less than 0.5 (Jayadas & Ambujam, 2021). The Disaster Resilience Place Model (DROP) framework adopted to study the inherent resilience index for Zimbabwe consists of community capital, economic, infrastructure and social and health as indicators (Mavhura et al., 2021). Access to basic services, assets, social safety nets, and adaptive capacity are the indicators used for developing micro-level resilience index. Natural disasters have a negative impact on education, crop variety, and consumption which is unavoidable as residents are unable to develop enough adopting abilities (Mondal et al., 2023). The study found that social capital (SC), ecological stability and livelihood resources such as cropland, livestock, livelihood diversification, and infrastructure are the main determinants impacting households' resilience to climate induced shocks (Asmamaw et al. (2019). The sustainable livelihood framework (SLF) was adopted to study the livelihood resilience in the mining sector. This framework consists of five capitals (Arhin et al. 2022). Resilience outcomes are the changes in a household's ability to respond to climate hazards.

With this background, the present study was undertaken with the objective of assessing the farmers' resilience towards natural disasters and developing a disaster management model to overcome the crisis effectively.

This framework highlights community social, political, and physical situations as well as the livelihoods that are impacted by trends, shocks, and seasonal changes from the outside world. Disaster resilience is related to the SLF approach, even though it primarily emphasizes the idea of sustainability (Tobin, 1999). Five dimensions make up the resilience tool. There are distinct indicators for every dimension. A farmer is therefore said to be more resilient if he performs on average better in each of these five areas. Overall, the farmers' resilience framework consists of 15 indicators given in Table 1. For example, road connectivity is an indicator of the ‘physical dimension’. Better road connectivity leads to faster and more efficient displacement of people from affected areas. Therefore, farmers are more likely to be resilient to shocks.

For the purpose of determining the weights, principal component analysis (PCA) was performed to determine the weightage for each indicator or variable under five dimensions. Based on component loadings of the first principal component with eigenvalue > 1 and a higher percentage of variance, the respondent score has been weighted (Filmer & Pritchett, 2001; Jayadas & Ambujam, 2021). The factor loadings of the first principal component are given in Table 2.

The various resilient indices have been computed, and the indices have been quantized. Table 3 provides the quantile-based classification of farmer resilience.

The resilience index has been calculated for the five dimensions, namely economic, social, human, physical and natural, followed by the FRI for both districts together. The factor loadings of the first principal component along with the eigenvalues and cumulative variance are given in Table 2 for fixing the weightage of each variable.

The overall FRI value for the study area is 0.55. Financial capital (FRI value = 0.641) has been perceived as the relatively most important indicator contributing to the resilience of a farmer. Social (0.612) and physical capitals (0.601) also equally contributed whereas the index value of human (0.496) and natural capitals (0.401) is less than the overall FRI score which means that those two indicators are perceived as least important by the farmer. The resilience index values for all capital are high except natural capital which is 0.401 because in the short run, farmers cannot cope with the effects of natural disasters and it takes time for them to recover from the damage caused to soil and land productivity. Farmers in both districts are financially better and have alternate source of income. It was also found that most of the farmers are semi-medium farmers having a landholding more than 2–4 ha. This might be the reason for the high financial capital index (0.641) which is shown in Figure 3. The findings are in line with Jayadas & Ambujam (2021).

After calculating the different resilient indices, the indices have been grouped into quantiles. The classification of farmer resilience based on the quantile is shown in Table 3.

Table 3 represents the distribution of respondents based on resilience. It was found that nearly half (39.58%) of the respondents had high resilience towards natural disasters followed by medium (34.17%) and low (26.25%), respectively. The results reveal that the majority of the respondents have high resilience towards natural disasters. This is because a majority of the farmers are getting crop insurance from the government and they are practising crop diversification with a range of 0.4–0.6 as a resilient strategy. In addition, farmers in the study area often possess traditional knowledge passed down through generations. This knowledge incorporates practices that have proven effective in managing agricultural risks. This combination of traditional wisdom and innovation allows them to find solutions to challenges posed by natural disasters.

At the same time, these farmers most of the time do not wait for external interventions and develop their own adaptation strategies to cope with disasters. In many cases, there is a good understanding of the challenges and problems faced by the farmers, they also know which strategy to adopt in order to tackle. However, in some cases, they lacked the capacity to implement the necessary changes due to lack of technical knowledge and fatalistic behaviour, which impedes the implementation of resilient strategies. Results are consistent with the results of Reddy et al. (2015).

SEM is an advanced statistical technique to study mainly social phenomenon, processes or behaviour by estimating the relationship between latent constructs and their indicator variables. The same was employed here to confirm the theoretical framework developed for the farmer resilience index based on a sustainable livelihood framework.

The internal consistency and reliability were evaluated using Cronbach's alpha (>0.7) for each dimension used for the development of the FRI. After calculating internal consistency, two-step approach of SEM comprising of measurement model and structural model was followed.

The measurement model is developed based on 1. CFA and 2. path diagram obtained is the structural model which is depicted in Figure 5.

Cronbach's alpha reliability was employed to identify the internal reliability indices of each construct using SPSS version 26. Table 4 indicates that all the constructs are highly reliable (>0.70) as suggested by George & Mallery (2010) in the model.

CFA was performed with 240 respondents to determine how well the data fit the scale model that was generated in which four measures were assessed to indicate the overall fitness of the measurement model. The four metrics were P-value, comparative fit index (CFI), Tucker–Lewis index (TLI) and incremental fit index(IFI), The results from CFA revealed that the model has acceptable fit measures given by Hu & Bentler (1999), shown in Table 5.

Physical capital (PC) indicator variables: PC1 (Possession of transportation), PC2 (Access to electricity) and PC3 (Road connectivity)

Social capital (SC) indicator variables: SC1 (Extent of social cohesiveness), SC2 (Use of social networking sites and SC3 (Urban connectivity)

Economic capital (EC) indicator variables: EC1 (Access to credit), EC2 (Financial safety) and EC3 (saving part of income)

Human capital (HC) indicator variables: HC1 (Participation in agricultural extension activity), HC2 (Health status) and HC3 (No of earning members)

Natural capital (HC) indicator variables: NC1 (Land productivity), NC2 (Water availability) and NC3 (Soil fertility status)

Overall resilience score : RS1, RS2, RS3, RS4 and RS5

The structural model was constructed based on the results from the CFA. The assessment of parameters and goodness of fit (GOF) proved that the structural model is good (P = 0.000, CFI = 0.838; IFI = 0.839; TLI = 0.814). Table 7 shows the model fit measures for SEM.

Figure 5 portrays the statistically significant estimates and casual paths (hypotheses).

The SEM analysis revealed that the latent variable resilience of the farmers showed a significant positive association with four (physical, social, financial and natural) capitals except HC which shows a negative significant association with resilience. The PC (β = 0.206, p = <0.05), social (β = 0.131, p = <0.001), financial (β = 0.552, p = <0.05), natural (β = 0.184, p = <0.05) and HC (β = −0.435, p = <0.05) were found significant at 5, 10, 5, 5 and 5% significance (α) levels, respectively. This revealed that any effort to improve these four dimensions would undoubtedly enhance the resilience of the farmers in the two districts. The SEM SPSS AMOS output expressed the relationship between the latent variables – resilience and components which are summarized in Table 8.

The SEM results indicated that four (physical, social, financial and natural) capitals are found to have a positive and significant relationship with resilient score. Hence improvement in those four capitals significantly improves the resilience of farmers' towards natural disasters.

From Table 8, it was found that HC shows negative and significant relationship (∝ = −0.435, p = 0.007) with resilient score at 1% level of probability. This is due to local farmers not participating in much training programmes regarding disaster coping mechanisms such as disaster preparedness or disaster recovery. This leads to lack of knowledge and skills necessary to adopt modern agricultural practices, manage risks effectively, and make informed decisions. Without adequate knowledge and skills, farmers may struggle to respond and adapt to challenges posed by natural disasters. Lack of knowledge and skill might be the reason for having a negative significant association of HC with resilience.

From Table 8, it was found that SC shows a positive and significant relationship (∝ = 0.131, p = 0.005) with a resilient score at a 1% level of probability. It was found that SC allows farmers in the study area to access valuable information and knowledge through social networks. They can learn from each other's experiences, and exchange information about climate patterns, innovative farming practices, and effective risk management strategies. This enhances their ability to adapt to changing conditions and build resilience. Due to this reason, SC has a positive and significant relationship with resilience score.

From Table 8, it was found that financial capital shows a positive and significant relationship (∝ = 0.552, p = 0.020) with a resilient score at a 5% level of probability. Similarly, savings are one of the important factors in identifying EC. In the study area, more than half of households have high savings through alternate sources of income. In addition, farmers also have better access to credit from institutional sources which resulted in an improvement in resilience among farmers towards disasters. This is the reason behind positive significant association of financial capital with resilience.

From Table 8, it was found that PC shows a positive and significant relationship (∝ = 0.206, p = 0.052) with a resilient score at a 5% level of probability. Infrastructure facilities like electricity and transportation are good in the study area. This helps in quick displacement from disaster-prone areas to less vulnerable areas. Hence, PC has a positive and significant relationship with resilience score.

From Table 8, it was found that PC shows a positive and significant relationship (∝ = 0.184, p = 0.060) with a resilient score at 5% level of probability. Although the farmers have a less natural capital score, this capital is also significantly affecting their resilience. Since all capital significantly affects the farmers' resilience, it is important for the government to focus on all five capitals which will automatically improve the livelihood of individuals and also their resilience to natural disasters. Capacity building and training programmes in disaster coping mechanisms have to be initiated by the government officials which will help the farmers to cope up with future shocks.

FRI concluded that half of the respondents under the Krishna and NTR districts are resilient to natural disasters whereas the remaining half are less resilient. The districts that are more vulnerable should be given priority for introducing any interventions. Given the priority for investment, policymakers can start with the most vulnerable locations and work their way down, reducing the overall impact of disasters. Additionally, it is crucial that climate awareness and capacity building become an essential component of all development programmes. The study also showed that four (physical, social, financial and natural) capitals positively and significantly impacted the farmers' resilience towards natural disasters. The developed disaster model may stimulate the policymakers to direct their efforts towards taking effective and sustainable preventive measures before the disaster occurs and ensure prompt response, recovery, and reconstruction during and after the disaster so as to build India's overall disaster resilience.

Funding was provided by Indian Council of Social Science Research (ICSSR).

All relevant data are available from an online repository or repositories: https://docs.google.com/spreadsheets/d/1eiBJsmOOHq8LEs_5odrFT-NtxfFBwoMF/edit#gid=1161945972.

The authors declare there is no conflict.