The COVID-19 pandemic exacerbated global poverty, underscoring the need for enhanced sustainable livelihood resilience among farmers. While China successfully eradicated absolute poverty, maintaining these gains and addressing relative poverty remains crucial, particularly in vulnerable mountainous areas. Farmers in these regions face numerous risks, including those related to illness, education, and unstable employment, increasing their vulnerability to returning to poverty. This study addresses the gap in understanding and measuring farmers' sustainable livelihood resilience, which is crucial for long-term poverty alleviation efforts. The concept of resilience, originating in physics and ecology, is applied to the context of farmers' livelihoods, integrating it with the established sustainable livelihood framework. Previous research on household livelihood resilience has mainly focused on analyzing the response relationship of livelihoods around key variables or using traditional poverty frameworks for comprehensive evaluation. However, these studies often lack a comprehensive consideration of dynamic interactions within the livelihood system and between farmers and the environment, and quantitative analysis methods. This study aims to fill this gap by developing a comprehensive framework and quantitative model to measure and analyze farmers' sustainable livelihood resilience in Fugong County, a formerly impoverished region in Yunnan Province, China.

Existing literature on household livelihood resilience often analyzes the relationship between livelihood status and key variables or utilizes established poverty frameworks. Scholars have used various methods, including comprehensive evaluation, scenario simulation, and dynamic models, to assess resilience in different contexts, such as post-disaster reconstruction, urban resilience, and tourism. Common dimensions used to measure farmers' sustainable livelihood resilience include buffer capacity, adaptation capacity, and self-organization capacity. However, a consistent paradigm for measuring resilience is lacking, and studies often rely on qualitative descriptive analyses, neglecting the interaction mechanisms among resilience components. This study builds upon previous work by incorporating the concept of resilience into the sustainable livelihood framework and emphasizing the quantitative measurement and analysis of farmers' resilience, thereby integrating traditional livelihood analysis with resilience theory.

This study uses Fugong County, Yunnan Province, as a case study area. Data collected include digital elevation models, Landsat 8 imagery, infrastructure data, administrative boundaries, road networks, poverty alleviation data (2015-2018), and socioeconomic data. Data preprocessing included cleaning, image processing, projection transformation, georeferencing, and topological inspections. The study integrates sustainable livelihood theory and resilience theory to develop a descriptive framework for farmers' sustainable livelihood resilience. This framework consists of three core elements: livelihood risks (natural disasters, policy changes, disease, etc.), the farmers' livelihood system (natural, physical, human, social, and financial capital, and internal and external disturbances), and farmers' livelihood resilience (buffer capacity, self-organization capacity, and learning capacity). An index system is constructed to measure farmers' sustainable livelihood resilience based on these three dimensions. A cloud-model-based multi-level fuzzy comprehensive evaluation model is employed to quantify the resilience levels. The entropy weight method is used to determine the weight of each indicator. A nine-point Likert scale is used to assess the importance of each indicator, and these values are converted into cloud model importance scales (expectation, entropy, and super-entropy). A fuzzy comprehensive evaluation is conducted using these scales, and the final score is calculated using a formula that combines expectation, entropy, and super-entropy with weights. The coupling coordination degree model is applied to analyze the interaction and coordinated development among the three resilience dimensions. A decision tree method is then used to classify farmers into different resilience state types based on their scores and coupling coordination degree.

The study reveals a generally upward trend in farmers' sustainable livelihood resilience in Fugong County from 2015 to 2018, although the increase was modest. Buffer capacity showed significant improvement, while self-organization capacity and learning capacity remained relatively stable. Spatial distribution of resilience was heterogeneous, with higher levels concentrated in the central region and lower levels in the southwest and northwest. At the village level, Shangpa, Shidi, and Chisadi villages consistently exhibited high resilience, while Dadake, Wawa, and Mujiajia villages showed consistently low resilience. The coupling coordination degree also showed an upward trend, with higher levels in the central region and lower levels in the north and south. The decision tree analysis classified villages into six resilience states: stable promotion, benign promotion, stagnation, mild recession, severe recession, and chaotic period. The proportion of villages in the promotion categories increased, while those in the recession categories decreased. The proportion of villages in the stagnation and chaotic categories remained substantial. The analysis suggests that the three dimensions of buffer capacity, self-organization capacity, and learning capacity interact synergistically, and the lack of one dimension negatively impacts overall resilience.

The findings highlight the importance of a multidimensional approach to measuring and enhancing farmers' sustainable livelihood resilience. The observed improvement in buffer capacity suggests that economic development strategies have had some success. However, the relatively stable self-organization and learning capacities indicate a need to improve social infrastructure and access to information and skills development. The heterogeneous spatial distribution of resilience emphasizes the importance of regionally tailored policies and interventions. The classification of villages into different resilience states allows for targeted support programs based on specific needs and challenges. The synergistic relationship among the three resilience dimensions points towards integrated approaches to promoting sustainable livelihoods that address economic, social, and knowledge-based aspects. The substantial number of villages remaining in stagnation or recession states shows that there is still significant work to be done to ensure long-term poverty reduction and sustainable development.

This study provides a comprehensive framework and quantitative model for measuring farmers' sustainable livelihood resilience. The findings reveal spatial and temporal variations in resilience, highlighting the importance of targeted interventions. Future research could extend the study period to include the impact of the COVID-19 pandemic, analyze the influence of various factors on resilience changes, and investigate the underlying mechanisms within the framework in more depth. The results offer valuable insights for policymakers in designing effective poverty alleviation and rural development strategies.

The study's timeframe (2015-2018) limits the analysis of long-term changes and the impact of recent events like the COVID-19 pandemic. Detailed analysis of how various influencing factors affect resilience levels was also constrained by space limitations. Further research is needed to address these limitations for a more complete understanding of the dynamics of farmers' sustainable livelihood resilience.